Citation
Subcellular localization and trafficking of amino acid transporters

Material Information

Title:
Subcellular localization and trafficking of amino acid transporters
Creator:
McDonald, Kelly Kristin, 1971-
Publication Date:
Language:
English
Physical Description:
x, 185 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Amino acids ( jstor )
Antibodies ( jstor )
Caveolins ( jstor )
Cell membranes ( jstor )
Fibroblasts ( jstor )
Laboratory staining techniques ( jstor )
Lysosomes ( jstor )
Microscopy ( jstor )
Proteins ( jstor )
Rabbits ( jstor )
Amino Acids -- metabolism ( mesh )
Carrier Proteins ( mesh )
Cell Membrane ( mesh )
Department of Biochemistry and Molecular Biology thesis Ph.D ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Biochemistry and Molecular Biology -- UF ( mesh )
Nuclear Envelope ( mesh )
Protein Binding ( mesh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 173-184).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Kelly Kristin McDonald.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
029482961 ( ALEPH )
51619876 ( OCLC )

Downloads

This item has the following downloads:


Full Text











SUBCELLULAR LOCALIZATION AND TRAFFICKING
OF AMINO ACID TRANSPORTERS











By

KELLY KRISTIN MCDONALD


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1998














ACKNOWLEDGMENTS


I would like to thank the members of my supervisory committee: Drs. Brian Cain,

William Dunn, Susan Frost, Mike Kilberg, and Peter McGuire. I wish to extend special

thanks to my mentor, Mike Kilberg, as well as Dr. Edward Block for his contribution to

the work discussed in Chapter 5. I would also like to acknowledge Dr. Stephen Wang for

his valuable instruction on the deconvolution microscope, David Parks, in the Center for

Structure Biology Computer Core, and Stephen Nowicki for valuable support and

friendship. Lastly, I would like to thank my parents, Drs. Maurice and Patricia

McDonald for their guidance, encouragement, and love.














TABLE OF CONTENTS

A CKN O W LED GM EN TS ................................................................................................. ii

LIST OF TA BLES............................................................................................................. v

LIST OF FIGU RES .......................................................................................................... vi

A BSTRA CT...................................................................................................................... ix

CHA PTER 1 IN TROD U CTION ...................................................................................... 1

O verview of M am m alian A m ino A cid Transport................................................ 1
Trafficking of M em brane Proteins..................................................................... 11
Cell Biological Techniques for Studying Protein Trafficking........................... 15

CHAPTER 2 MATERIALS AND METHODS ............................................................. 17

M materials ............................................................................................................ 17
M ethods.............................................................................................................. 18

CHAPTER 3 DISTRIBUTION OF THE GLUTAMATE TRANSPORTERS.............. 28

Introduction........................................................................................................ 28
M ethods.............................................................................................................. 36
Results................................................................................................................ 39
D iscussion.......................................................................................................... 48

CHAPTER 4 LYSINURIC PROTEIN INTOLERANCE.............................................. 69

Introduction........................................................................................................ 69
Results................................................................................................................ 76
D iscussion.......................................................................................................... 89

CHAPTER 5 CAVEOLAR COMPLEX BETWEEN THE CATIONIC AMINO
ACID TRANSPORTER 1 AND ENDOTHELIAL NITRIC
OX ID E SYN TH A SE .............................................................................. 124

Introduction...................................................................................................... 124
M ethods............................................................................................................ 129
Results.............................................................................................................. 132

iii








Discussion........................................................................................................ 143

CHAPTER 6 CONCLUSIONS AND FUTURE DIRECTIONS ................................. 169

LITERATURE CITED .................................................................................................. 173

BIOGRAPHICAL SKETCH ......................................................................................... 185












































iv














LIST OF TABLES


Table p e

1-1. cDNA Clones of Amino Acid Transporters in CAT and EAAT Families.................. 2

3-1. Antibodies against Glutamate Transporters............................................................... 39

4-1. Antibodies for Immunofluorescence Studies............................................................. 78

5-1. Antibodies for Immunofluorescence Studies........................................................... 130

5-2. Immunodepletion of CAT I-mediated Arginine Transport Activity
by anti-eN O S A ntibody .......................................................................................... 139














LIST OF FIGURES


Figure pgAe

3-1. Extracellular staining of human fibroblasts with EAAT3 antibody...................... 53

3-2. Intracellular staining of human fibroblasts with EAAT3 antibody
and co-localization with organelle-specific antibodies.......................................... 55

3-3. Nuclear staining of human fibroblasts with EAAT1 -R antibody and
co-localization with nucleus-specific antibodies ................................................... 57

3-4. Intracellular staining of human fibroblasts with EAAT1-S and
EA A T 1-C antibodies ............................................................................................. 59

3-5. Intracellular staining of Hela cells with EAAT1-R and EAAT1-S
antibodies ............................................................................................................... 6 1

3-6. Immunoblot analysis of EAAT1 in the nuclear and intracellular
membrane fractions from human fibroblasts......................................................... 62

3-7. Expression of GFP and GFP-EAAT 1 fusion proteins in human
fi broblasts............................................................................................................... 64

3-8. EAAT1 immunofluorescent staining of human fibroblasts
transfected with EAAT1-GFP(N3)........................................................................ 66

3-9. EAAT1 immunofluorescent staining of PAEC transfected with
EA A T 1 -G FP(N 3)................................................................................................... 68

4-1. Morphology of normal and LPI fibroblasts by light microscopy .......................... 97

4-2. Morphology of normal and LPI fibroblasts by electron microscopy..................... 99

4-3. Intracellular staining of normal and LPI human fibroblasts with the
C A T 1 antibody..................................................................................................... 10 1

4-4. Expression of GFP and the GFP(C3)-CAT1 fusion protein in
norm al hum an fibroblasts .................................................................................... 103

4-5. Expression of the GFP(C3)-CAT1 fusion protein in normal and
LPI hum an fibroblasts.......................................................................................... 105

vi








4-6. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the nuclear membrane and nucleolus .................................... 107

4-7. Intracellular staining of normal and LPI fibroblasts with antibodies
against plasma membrane and cytoskeletal proteins ........................................... 109

4-8. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the endoplasmic reticulum and Golgi Complex.................... 111

4-9. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the endocytic and recycling pathways................................... 113

4-10. Intracellular staining of normal and LPI fibroblasts with an
antibody against a lysosomal enzyme.................................................................. 115

4-11. Intracellular staining of normal and LPI fibroblasts with an
antibody against a lysosomal membrane protein................................................. 117

4-12. Visualization of acidic compartments of normal and LPI
fibroblasts with acridine orange........................................................................... 119

4-13. Lysosomal detection in normal and LPI fibroblasts following
chloroquine treatm ent .......................................................................................... 121

4-14. Lysosomal staining of normal and LPI cells expressing the
G FP(C3)-CA T 1 fusion protein ............................................................................ 123

5-1. Surface labeling of PAEC with the CAT 1 transporter antibody.......................... 149

5-2. Co-localization of CAT1 and caveolin on PAEC ............................................... 151

5-3. Co-localization of CAT1 and eNOS on PAEC.................................................... 153

5-4. Detection of CAT1 and eNOS in the Golgi of PAEC ......................................... 155

5-5. Expression of the GFP(C3)-CAT1 fusion protein in PAEC................................ 157

5-6. Immunofluorescent staining of PAEC transfected with GFP(C3)-
C A T 1 fusion protein ............................................................................................ 159

5-7. Distruption of CAT1/eNOS co-localization in PAEC treated with
nocodazole ........................................................................................................... 16 1

5-8. CAT1 immunoprecipitation of eNOS from solubilized PAEC
plasm a m em brane vesicles................................................................................... 162








5-9. Immunofluorescent staining of PAEC transfected with the
CATMUT1 palmitoylation mutant...................................................................... 164

5-10. Immunofluorescent staining of PAEC transfected with the
CATMUT3 palmitoylation mutant...................................................................... 166

5-11. eNOS staining of PAEC treated with varying concentrations of
extracellular L-arginine........................................................................................ 168














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

SUBCELLULAR LOCALIZATION AND TRAFFICKING
OF AMINO ACID TRANSPORTERS

By

Kelly Kristin McDonald

August, 1998


Chairman: Dr. Michael S. Kilberg
Major Department: Biochemistry and Molecular Biology

Mammalian amino acid transporters have been well-characterized with regard to

substrate specificity, kinetic parameters, and metabolic regulation. However, little

information is known concerning the cell biology or "life cycle" of plasma membrane

amino acid transporters. In this study, molecular and cell biological techniques, including

immunohistochemistry, transporter mutants and expression of green fluorescent protein

fusions, and deconvolution microscopy, were used to examine the cellular localization

and trafficking of specific amino acid transporters under normal and diseased conditions.

The availability of cDNAs and antibodies for specific members of the EAAT anionic and

the CAT cationic amino acid transporter families have provided an avenue for comparing

the subcellular distribution of amino acid transporters from the same, as well as different

gene families. Although similar in structure and function, the EAAT1 glutamate

transporter was detected primarily in the nuclear membrane in certain cell types, such as

ix








human fibroblasts, whereas the EAAT3 transporter was observed in intracellular vesicle

compartments and in concentrated clusters on the plasma membrane. Lysinuric Protein

Intolerance (LPI) provided a model system for investigating the subcellular organization

and trafficking pathways in a disease with visual morphological defects. A survey of

organelle integrity led to the discovery of an abnormal population of lysosomes in the

LPI fibroblasts. Future studies will investigate the contents of the LPI lysosomes, as well

as the amino acid transport across lysosome-enriched membrane preparations.

Immunostaining of porcine aortic endothelial cells (PAEC) revealed "patches" of the

cationic amino acid transporter CAT1 that co-localized with antibodies against caveolin

and endothelial nitric oxide synthase (eNOS). When incubated with solubilized PAEC

plasma membrane proteins, an eNOS-specific antibody immunoprecipitated CAT1-

specific arginine transport activity. These results document the existence of a caveolar

complex between CAT1 and eNOS in PAEC that provides a potential mechanism for the

efficient delivery of the arginine substrate to eNOS for nitric oxide (NO) production. The

individual projects presented in this thesis share a common goal in documenting the

cellular localization, and when possible, understanding the functional consequences, of

the transporter protein distribution.














CHAPTER 1
INTRODUCTION

Overview of Mammalian Amino Acid Transport

Over the past three decades, a variety of mammalian amino acid transport systems

have been characterized extensively for cell and substrate specificity, kinetic parameters,

and metabolic regulation. However, the original studies began with the pioneering work

of Van Slyke and Meyer in 1913, when they demonstrated that tissues accumulated

amino acids against a concentration gradient (Van Slyke and Meyer, 1913). In the early

1960s, the Christensen laboratory began to define specific transport systems that mediate

the flux of amino acids across the membrane bilayer based on the conformation, size, and

chemical properties of the amino acid side chain (Oxender and Christensen, 1963).

Rigorous investigation revealed that each system recognizes more than one amino acid,

distinct systems exhibit some degree of overlapping substrate reactivity, and various

amino acids are transported by more than one system. These observations were further

complicated by the discovery that distinct cell and tissue types express a different

combination of systems that work in concert to provide specific nutritional requirements.

The identification of individual proteins responsible for amino acid transport activity has

been complicated by the relatively low abundance of these proteins, as well as the

technical difficulty involved in isolating and purifying integral membrane proteins.

However, the advances in molecular biology over the last ten years have resulted in the









cloning and expression of more than 20 cDNAs encoding amino acid transporters. Based

on sequence homology, a number of the cloned transporters have been classified

according to gene families. Those related to the studies presented in this thesis are

summarized in Table 1-1.

Table 1-1
cDNA Clones of Amino Acid Transporters in CAT and EAAT Families
Alternate Deduced amino Substrate
Clone names acid length specificity Ions coupled

CAT 1 -- 622-624 cationic --
CAT2 CAT2a 658 cationic --
CAT2a CAT2b 659 cationic --
CAT3 619 cationic --
EAAT1 GLAST1 543 D,L-aspartate Nan, K-ou,
GluT L-glutamate Win
EAAT2 GLT1, GLTR 573 D,L-aspartate Na+i., K+o.u
GLAST2 L-glutamate Htm
EAAT3 EAACI1 523-525 D,L-aspartate Na1,, K+o,,
L-glutamate HWn
EAAT4 564 D,L-aspartate Nati,, K+ou,
L-glutamate HWn
EAAT5 560 D,L-aspartate Na+,, K+ou,
L-glutamate H+,,,


The "CAT family" is composed of four known transporters that mediate the Na-

independent transport of the cationic amino acids arginine, lysine, ornithine, and histidine

when positively charged (reviewed by MacLeod et al., 1994; Malandro and Kilberg,

1996). Although the members of this family share a common substrate specificity and

significant amino acid similarity, they differ in tissue distribution, affinity for substrate,

and most likely the functional role they play in metabolism. The "anionic transporter

family" currently contains five members that are responsible for the Na+-dependent

transport of the anionic amino acids glutamate and aspartate (reviewed by Malandro and









Kilberg, 1996). and are related to two members of the "ASCT family" that mediate Na'-

dependent transport of selected zwitterionic amino acids (Arriza et al., 1993; Shafqat et

al.. 1993; Liao and Lane, 1995). The Na/Cl'-dependent proline transporter, as well as all

four glycine transporting members of the "GLYT family," are components of a large

superfamily ofneurotransmitter transporters and are primarily expressed in the central

nervous system (reviewed by Malandro and Kilberg, 1996). Two cDNAs, designated

NBAT and 4F2hc, express proteins with only one to four transmembrane spanning

domains and comprise a unique family responsible for cationic and zwitterionic amino

acid transport (reviewed by Palacin, 1994; Malandro and Kilberg, 1996). However, it has

not yet been determined whether these proteins are actual transporters themselves, or

rather function as regulators or accessory subunits of a transporter complex.


Cationic Amino Acid Transport Systems

System y', the primary mechanism for the transport of cationic amino acids, was

first described by White and Christensen in the early 1980s (White and Christensen,

1982; White et al., 1982). This "transport activity," first characterized in the Ehrlich cell,

was shown to be Na'-independent, pH-insensitive, and stereoselective for L-amino acids

(reviewed by Kilberg et al., 1993). The tissue distribution was widespread, yet not

ubiquitous, and amino acids accumulated against a concentration gradient in response to

membrane potential. The uptake of amino acids was also subject to trans-stimulation, the

property of increased amino acid transport when substrate concentrations are elevated on

the opposite side of the membrane. Lastly, zwitterionic amino acids in the presence of

Na' ions could competitively inhibit the transport of cationic amino acids by System y"








(reviewed by Kilberg et al., 1993). Presently, four distinct proteins that specifically

mediate plasma membrane cationic amino acid transport have been cloned, and

demonstrate similar, yet not identical, properties to those of System y'.

I have chosen to study the primary cationic amino acid transporter, designated

CAT1, as well as several members of the glutamate family (described below). In 1989,

Cunningham and coworkers identified a cDNA clone that encoded the ecotropic murine

leukemia virus receptor (Albritton et al., 1989). It was later determined that the receptor

also functioned as a Na-independent, high-affinity transporter for arginine, lysine,

ornithine, and histidine when positively charged (Kim et al., 1991; Wang et al., 1991).

The corresponding human cDNA was cloned by Meruelo and coworkers (Yoshimoto et

al., 1991) and is 88% homologous to the murine transporter, but lacks homology in the

region of viral binding. This is consistent with the observation that the ecotropic murine

leukemia retrovirus is unable to infect human cells (Albritton et al., 1993). The human

CAT 1 sequence revealed a 629-amino acid protein with 12-14 predicted transmembrane

spanning domains. The CAT1 transporter mRNA is not expressed in the liver, but

otherwise appears to be ubiquitous (for review see Malandro and Kilberg, 1996).

Immunohistochemistry from our laboratory, using an antibody against the third

extracellular loop of the murine cDNA sequence, confirmed the lack of expression in rat

liver (Woodard et al., 1994).

MacLeod and coworkers cloned a second member of the CAT family (MacLeod

et al., 1990) from murine T cell lymphocytes (Kekuda et al., 1993). This protein,

originally called the Tea gene, was discovered during a search for genes involved in T

cell activation, however, its function as a transporter was suggested by its extensive








sequence homology to CAT1. The Tea gene, later renamed murine CAT2, is 61%

identical to CAT1 at the amino acid level, has 12 predicted membrane spanning domains,

and mediates Na'-independent high-affinity cationic amino acid transport in activated T

lymphocytes (for review see Malandro and Kilberg, 1996). Chimeric constructs of the

mouse and human CAT 1 sequences were used to identify the region of viral binding in

the third extracellular loop of the murine CAT1 protein (Albritton et al., 1993). The

CAT2 sequence, like the human CAT1 sequence, is divergent in this region and does not

function as a binding site for the retrovirus. Nitric oxide (NO) has been implicated in T

cell signaling by autocrine/paracrine pathways, and it has been proposed that the

expression of the CAT2 transporter during T cell activation may be related to the need for

arginine in the production of NO (MacLeod et al., 1994). Using the murine CAT2

sequence as a probe for screening a mouse liver cDNA library, Cunningham and

coworkers identified a third liver-specific member of the CAT family, murine CAT2a.

This protein is encoded by the same gene as CAT2, but as a result of differential splicing,

CAT2a has an additional stretch of 41 amino acid residues (358-398) between the eighth

and ninth membrane spanning domains. Despite the similarity in sequence, CAT2a

exhibits a 10-fold lower affinity for arginine than either CAT1 or CAT2 (Closs et al.,

1993). This kinetic difference suggests that the 41 amino acid region may be involved in

binding the amino acid during its translocation across the membrane.

Screening a rat brain cDNA library with probes designed from the murine CAT1

sequence isolated the most recent addition to the CAT family, CAT3 (Hosokawa et al.,

1997). The protein encoded by the rat CAT3 is comprised of 619 amino acids and shares

53-58% identity with the CAT family members previously described. The CAT3 protein









mediates the high-affinity, Na'-independent transport of cationic amino acids and shares

the greatest homology with the CAT1 family member. In the same year, the mouse

CAT3 cDNA was identified by Ito and Groudine during an attempt to isolate germ-layer

specific transcripts from mouse embryos (Ito and Groudine, 1997). The mouse CAT3

was localized to the brain by in situ hybridization studies, and exhibited the same

structural and transport characteristics as the rat homolog. The highly conserved tissue-

specificity between the rat and mouse proteins suggests an important role for CAT3 in the

brain. Detection of mRNA in brain capillaries suggests a role for CAT3 in the transport

of cationic amino acids by endothelial cells at the blood-brain barrier (Hosokawa et al.,

1997). Just as CAT2 may provide the inducible nitric oxide synthase (iNOS) with

substrate for NO production in T cells, it is feasible that CAT3 provides the neuronal

nitric oxide synthase (nNOS) isoform with the arginine required for NO production in the

nervous system.

The CAT family, comprising four functionally similar yet distinct proteins, is

only one of two families responsible for the transport of cationic amino acids. The other

family has been identified based on mRNA expression in Xenopus oocytes and currently

includes two proteins, NBAT and 4F2hc. Expression of either in oocytes induces Na'-

independent transport of cationic amino acids, but NBAT mediates the uptake of Na'-

independent zwitterionic amino acids whereas 4F2hc-catalyzed uptake of these substrates

is Na'-dependent (Bertran et al., 1992; Wells et al., 1992). These specificities correspond

to two known transport activities, called System b0' (NBAT) and System y*L (4F2hc),

respectively. The hydropathy plots of both transporters predict either one or four

membrane spanning domains. This is an interesting feature for two proteins that are









believed to function in transport, because most of the cloned transporters are thought to

span the membrane 12-14 times. That NBAT and 4F2hc have four or less trans-

membrane domains has led to the hypothesis that they may serve as modulators, or

comprise only one subunit of a multimeric transporter complex (for review see Palacin,

1994). Co-precipitation and cross-linking experiments have recently provided strong

evidence that NBAT, a 90 kDa protein, is associated with a 50 kDa protein (Wang and

Tate, 1995). The NBAT protein is expressed in the microvilli of proximal tubules of the

kidney and the mucosa of the small intestine (Kanai et al., 1992). Expression of NBAT

in Xenopus oocytes has been shown to induce the high-affinity uptake of cystine. This

finding provided key evidence for NBAT's involvement in cystinuria, an autosomal

recessive disorder characterized by the hyperexcretion of cystine and cationic amino acids

into the urine (Segal and Thier, 1989; reviewed by Palacin, 1994). At least six distinct

missense mutations in NBAT have been documented in different patients with the

disease, but all result in defective transport of cystine through the epithelial cells of the

renal tubule and intestinal tract (Calonge et al., 1994).

Anionic Amino Acid Transport Systems

The anionic transporter family currently includes at least five members that

mediate glutamate/aspartate transport. These Na-dependent transporters share a similar

structure with six predicted trans-membrane spanning domains in the N-terminal portions

and a large hydrophobic region at the C-termini that may represent additional trans-

membrane domains. Each transporter has at least two putative glycosylation sites and

shares from 40-68% amino acid sequence identity with the other members of the family.









Glutamate transport studies in salamander retina glial cells demonstrated that glutamate

uptake is electrogenic and coupled to the co-transport of three Na' ions and a H', as well

as the counter-transport of one K' ion (Zerangue and Kavanaugh, 1996). In the central

nervous system, the stoichiometry of the glutamate and ion transport must be tightly

regulated. Ischemic conditions following a stroke may lead to the breakdown of

electrochemical gradients as a result of lower ATP levels and reduced functioning of

Na',K ATPase proteins (Szatkowski and Attwell, 1994). If the ion gradients are

disrupted, then it is believed that the glutamate transporters can function in reverse,

resulting in the release of glutamate into the synaptic cleft, and subsequent neurotoxicity

and neuronal death (Kanai et al., 1995).

In an attempt to isolate galactosyltransferase from rat brain, Storck and coworkers

co-purified the first glutamate transporter, designated GLAST1 (Storck et al., 1992). The

isolated protein showed homology to the previously cloned bacterial glutamate and

monocarboxylate transporters, and its function as a glutamate transporter was confirmed

by expression in Xenopus oocytes followed by radiolabeled amino acid uptake (Klockner

et al., 1993). Pines and coworkers cloned the second glutamate transporter, GLT1, by

screening a rat cDNA expression library with antibodies generated against the partially

purified protein (Pines et al., 1992). Northern analysis and in situ hybridization detects

GLAST1 and GLT1 mRNA expression primarily in glial cells of the central nervous

system (Storck et al., 1992; Otori et al., 1994), where they play an important role in

clearing toxic levels of glutamate from the synaptic clefts. Although the mechanism is

unclear, recent data from Rothstein and coworkers indicate that abnormal GLT1 mRNA

species may be responsible for the decreased glutamate transport detected in patients with









Amyotrophic Lateral Sclerosis (ALS), but the defective glutamate transport is not thought

to be the principle cause of the disease (Lin et al., 1998). The third member of the

growing glutamate transporter family, EAAC 1, was cloned by oocyte expression cloning

using fractionated mRNA from rabbit intestine (Kanai and Hediger, 1992). Although

similar to GLAST1 and GLT1 in structure and transport properties, EAAC 1 expression is

neuron-specific in the brain. EAAC I transcripts have also been detected in small

intestine, kidney, liver, heart, placenta, and skeletal muscle (reviewed by Malandro and

Kilberg, 1996). In 1994, the human homolog to GLAST1 was identified (Arriza et al.,

1994; Kawakami et al., 1994) and given the name EAAT1, for Excitatory Amino Acid

Transporter 1. Shortly after, the human GLT1 sequence was reported (Arriza et al., 1994;

Manfras et al., 1994) and designated EAAT2, and the human homolog to EAAC1 was

cloned (Arriza et al., 1994; Kanai et al., 1995) and called EAAT3. It has since become

acceptable to refer to the glutamate transporters by the EAAT nomenclature regardless of

the species.

EAAT4 was isolated using degenerative oligonucleotide primers corresponding to

conserved sequences within the other members of the glutamate family (Fairman et al.,

1995). Northern analysis identified EAAT4 mRNA in the cerebellum and placenta, and

transport studies in oocytes demonstrated a high-affinity glutamate uptake that was

associated with chloride conductance. The final member of this family, EAAT5, was

cloned by Arriza et al. by screening a human retinal cDNA library with a glutamate

transporter cDNA isolated from salamander retina (Arriza et al., 1997). Like EAAT4,

EAAT5 may play a role in ion conductance instead of, or in addition to, providing

neurotransmitter clearance at the synaptic cleft. Electrophysiological studies of both








EAAT4 and EAAT5 have shown a large chloride conductance in addition to transport

activity. In the case of EAAT5, this associated chloride conductance may participate in

visual processing (Arriza et al., 1997). A potential PSD-95-binding motif was identified

in the C-terminus of EAAT5 (Arriza et al., 1997). PSD-95, a cytoskeleton-associated

synaptic protein, has been shown to bind to C-terminal sequences in both the N-methyl-

D-aspartate (NMDA) receptor and Shaker-type potassium channels (Cho et al., 1992).

ASCT Amino Acid Transport Systems

Two zwitterionic amino acid transporters, ASCT1 (Shafqat et al., 1993; Arriza et

al., 1993) and ASCT2 (Kekuda et al., 1996; Utsunomiya-Tate et al., 1996), are 56%

identical to one another at the amino acid level, and both are approximately 40% identical

to the five glutamate transporters. The "ASC" transporters are Na'-dependent, and

although they exhibit a broad substrate specificity, they prefer amino acids with

hydroxyl- or sulfydryl-containing side chains serinee, cysteine, and threonine). Like

System ASC, the transporters exhibit the property of trans-stimulation even though this

activity is typically a characteristic of Na-independent transporters. Northern blot

analysis revealed highest expression of ASCT1 in the brain, skeletal muscle, and pancreas

(Shafquat et al., 1993; Arriza et al., 1993). ASCT2, also known as ATB, mRNA is

expressed in lung, skeletal muscle, kidney, large intestine, testes, and adipose tissue

(Kekuda et al., 1996; Utsunomiya et al., 1996). This transporter, only recently cloned by

homology to ASCTI, has been shown to exhibit a broader substrate specificity than

ASCT1 (Kekuda et al., 1996; Utsunomiya et al., 1996). In fact, one of the distinguishing








features between the two is the acceptance of glutamine by ASCT2, but not by ASCT1

(Utsunomiya et al.. 1996).

ASCT1 exhibits a unique pH-dependent substrate specificity (Tamarappoo et al.,

1996), first described for System ASC by Christensen and colleagues (Vadgama and

Christensen, 1984). At neutral pH, ASCT1 preferentially transports zwitterionic amino

acids, whereas if the assay pH is lowered to 5.5, the transporter will accept both

zwitterionic and anionic amino acids. Transport assays using anionic amino acid analogs

to compete with the zwitterionic substrates at pH 5.5, indicate that the substrates may

share the same binding region (Tamarappoo et al., 1996). One hypothesis to explain this

pH effect is that one or more of the eight histidine residues in ASCT 1 accepts) a positive

charge when the pH is lowered below 6.0 (the pKa for the histidine side chain). The

positively charged histidine(s) may serve as a binding site for negatively charged amino

acids, such as glutamate, aspartate, and cysteate. ASCT2 also has a low affinity for

glutamate at neutral pH, and the affinity for the anionic amino acid increases as the assay

pH is lowered (Utsunomiya et al., 1996; Kekuda et al., 1996).


Trafficking of Membrane Proteins

The cloning and expression of a number of the amino acid transporters have led to

the generation of sequence-specific antibodies from corresponding peptides and fusion

proteins. The availability of antibodies and the development of high-resolution

microscopes have allowed investigators to initiate studies on the cell biology of amino

acid transporters, an area of the field in which little is known. The first mammalian

amino acid transporter antibody was produced in 1992. In recent years, many of the








individual steps that contribute to the "life cycle" of plasma membrane proteins have been

documented. Although little information is available, it is likely that many similarities

exist between the "life cycle" of amino acid transporters and other plasma membrane

proteins. From biogenesis to degradation, these proteins are transported through a

complex system of membrane compartments and organelles by specific vesicles that bud

from a donor membrane and fuse with a target membrane (Ivessa et al., 1995). A

combination of coat proteins and several classes of monomeric GTPases (i.e., Rabs) are

believed to regulate certain steps in vesicle trafficking. The SNARE hypothesis describes

the mechanism by which transport vesicles target membranes (Alberts et al., 1994). v-

SNAREs are proteins on the vesicle membranes and t-SNAREs reside on the target

membranes. v-SNAREs and t-SNAREs are suspected to function as structural proteins

that interact at the point of vesicle docking. Over 30 different Rab proteins have been

identified and each is believed to play some role to ensure the specificity of individual

vesicle docking/fusion events of the membrane trafficking pathways (Nuoffer and Balch,

1994).

Three primary trafficking pathways have been described for various membrane

proteins: the biosynthetic-secretory pathway, the endocytic-exocytic pathway, and the

degradative pathway. In the biosynthetic pathway, integral membrane proteins and

secretary proteins are co-translationally inserted into the membrane or the lumen,

respectively, of the endoplasmic reticulum (ER) where early oligosaccharide modification

and proteolytic processing begin. Select proteins receive fatty acylation moieties either

during or following, the translation event (reviewed by Solski et al., 1995). Myristic and

palmitic acid modifications contribute to membrane binding and stability, and more








recently, have been implicated in targeting certain proteins to plasma membrane caveolae

(Song et al., 1996). Following translation, proteins are packaged into vesicles and

transported to the Golgi apparatus for the completion of glycosylation and folding events.

The developing proteins are shuttled from the cis, to the medial, to the trans-Golgi

compartment, ultimately arriving at the trans-Golgi network (TGN) where they are sorted

according to their final destinations (Alberts et al., 1994).

Membrane proteins also participate in the endocytic/exocytic pathway. Several

distinct forms of endocytosis have been described (reviewed by Watts and Marsh, 1992;

Alberts et al., 1994). During pinocytosis, small invaginated plasma membrane vesicles of

less than, or equal to 150 nm in diameter, constituatively carry fluids and solutes into the

cell. Phagocytosis, on the other hand, results in the regulated ingestion of large particles

via plasma membrane derived vesicles of greater than 250 nm in diameter, and is

generally the responsibility of specialized cells. It is assumed that there is an

internalization of many plasma membrane proteins during these two processes. Most

animal cells take up specific macromolecules by a process called receptor-mediated

endocytosis. During this event, receptor-ligand complexes participating in this cycle are

internalized by clatherin-coated pits on the plasma membrane and delivered to early

endosomes. The acidic environment of the endosome results in the dissociation of the

ligands, which advance via late endosomes to ultimate lysosomal degradation. Some

receptors are recycled directly back to the plasma membrane from the endosomal

compartment, whereas others recycle by way of an intermediate step in the TGN. The

transferring receptor (TfR) interacts with iron-bound transferring at the plasma membrane

(reviewed by Hansen et al., 1993; Alberts et al., 1994). Following endocytosis, iron








molecules are released from the transferring in response to the low pH of the endosome.

This allows the unbound transferring to recycle to the plasma membrane with its receptor.

The transferring molecule is then freed so that it may sequester more extracellular iron (De

Silva et al., 1996).

The final trafficking pathway that contributes to the life cycle of membrane

proteins involves the transport of proteins to the lysosomes for degradation (reviewed by

Komrnfeld and Mellman, 1989), or the ubiquitination and subsequent degradation by

proteosomes (reviewed by Hershko and Ciechanover, 1992). Materials from multiple

pathways are emptied into the lysosomes, where acid hydrolases function in the regulated

digestion of macromolecules such as membrane proteins. The mannose-6-phosphate

receptor (M6PR) binds to the phosphorylated mannose modification on lysosomal-

destined digestive enzymes and shuttles these acid hydrolases from the TGN to the late

endosomes. The acid hydrolases eventually end up in lysosomes and the M6PR recycles

to the TGN. Therefore, the M6PR is an excellent marker for the late endosomal and TGN

compartments. A second trafficking pathway that leads to degradation evolves from

endocytosis. In a process that is poorly understood, materials destined for degradation

are transferred from early endosomes to late endosomes to lysosomes (Alberts et al.,

1994). Antibodies against mammalian amino acid transporters have become available

only in the last couple of years, so no information has been published concerning the

molecular mechanisms by which these transporters are degraded.











Cell Biological Techniques for Studying Protein Trafficking

Indirect immunofluorescence has provided a way of studying the subcellular

localization and trafficking of proteins using a combination of antibodies and trafficking

inhibitors. Indirect immunofluorescence is a sensitive method for detecting a protein of

interest because many molecules of the secondary antibody recognize each molecule of

primary antibody, which is raised against a specific peptide or protein. This results in an

amplification of the signal because the secondary antibody is covalently attached to a

fluorochrome that fluoresces when exposed to a specific wavelength of light. However,

problems may arise if an antibody cannot be raised against a desired protein, or if an

antibody produces a high level of background by cross-reacting with other cellular

proteins or artifacts. To avoid some problems commonly associated with antibodies, and

to investigate living cells, a new technique that utilizes the autofluorescence of the green

fluorescent protein (GFP) has replaced, or is being used in conjunction with, antibody-

labeling techniques.

GFP, a 27 kDa protein native to the bioluminescent jellyfish, Aequorea victoria,

produces a bright green color when stimulated by blue or UV light (reviewed by Steams,

1995). GFP expression is species-independent and can be introduced into prokaryotic

and eukaryotic cells without the requirement of specific cofactors, substrates, or

additional gene products. The GFP protein is small and globular, and in most cases does

not interfere with the synthesis, trafficking, or activity of the fusion protein product.

Whereas antibody labeling often involves the use of a fixative, GFP constructs can be









viewed in fixed or living cells. Certain GFP variants have been optimized for use in

mammalian cells and are now commercially available. These variants that contain

double-amino-acid substitutions Phe-64 to Leu and Ser-65 to Thr result in a 35-fold

increase in fluorescence over wild type GFP (Cormack et al., 1996). In addition,

expression has been enhanced by the introduction of silent mutations in the coding

sequence that correspond to human codon-usage preferences. The GFP serves as a

genetic tag that can be conveniently added to the protein coding sequence of a cDNA.

In this study, molecular and cell biological techniques, including

immunohistochemistry, GFP expression, and deconvolution microscopy, have been used

to examine specific aspects of the "life cycle" of amino acid transporters. The individual

projects presented in this thesis share a common goal in documenting the cellular

localization, and when possible, understanding the functional consequences, of

transporter distribution under normal and diseased conditions.














CHAPTER 2
MATERIALS AND METHODS

This Chapter contains the general materials and methods used during the course of

this research project. Chapter 3 and Chapter 5 include additional Methods Sections that

are specific for the work described in those chapters.


Materials

Fibronectin, bovine serum albumin (BSA), Triton X-100, and Triton X-l 14,

kanamycin, nocodozole, dimethyl sulfoxide, and adenosine triphosphate (ATP) were

purchased from Sigma Chemical Company (St. Louis, MO). RPMI and MEM media,

goat serum (NGS), fetal bovine serum (FBS), lipofectamine, OptiMEM, EcoRI and

Hindll restriction enzymes and buffers, T4 polynucleotide kinase and buffer, and all

PCR primers and mutagenesis oligonucleotides were purchased from Gibco BRL

(Gaithersburg, MD). Paraformaldehyde (PFA), glycine and methanol (MeOH) were

purchased from Fisher Scientific (Pittsburgh, PA). Fluoromount-G was purchased from

Southern Biotechnology Associates (Birmingham, AL). The MORPH mutagenesis kit

was obtained from 5 Prime -) 3 Prime (Boulder, CO), the pCR2.1 TA cloning kit was

obtained from InVitrogen (Carlsbad, CA), and the DNA gel purification kit and DNA

plasmid purification kits were obtained from Quiagen (Valencia, CA.). The

nitrocellulose membranes were purchased from Cuno, Inc. (Meridian, CT) and the

enhanced chemiluminesence reagents were purchased from Pierce (Rockford, IL). The









MCAT1 cDNA was a generous gift from Dr. James Cunningham at Brigham and

Women's Hospital (Boston, MA), and the rat EAAC I cDNA was cloned from a rat

hippocampal library (Velaz-Faircloth et al., 1996). All of the antibodies are described in

the chapters in which they were used.


Methods

Cell culture. Pulmonary artery endothelial cells (PAEC) were prepared by

collaborators in Dr. Edward Block's laboratory. PAEC were isolated by collagenase

treatment of the main pulmonary artery of 6-7-month-old pigs and were cultured for 3-7

passages as described by Block and coworkers (Block et al., 1989). One hundred-mm

dishes or 6-well trays were incubated with 7-10 pg/ml of fibronectin (Sigma Chemical

Co.), dissolved in RPMI medium, overnight at 37C under a humidified atmosphere of

5% CO,-95% air. Prior to plating cells, the fibronectin solution was aspirated, and

dishes/trays were allowed to dry for 30 min in a culture hood under a UV lamp. Once

plated, PAEC were maintained in RPMI + 4% or 10% fetal bovine serum (FBS). Hela,

HepG2 (human hepatoma), HEK 293 (human embryonic kidney), CHO and BNL.CL2

(mouse hepatocytes) cells were maintained in Eagle's minimal medium (MEM) + 4%

FBS. Cultured fibroblasts from Finnish LPI patients and sex-age-matched normal

controls were obtained from Dr. Olli Simell at the Central Hospital, University of Turku

(Turku, Finland). The fibroblasts were cultured in 75-mm flasks and maintained in

MEM, supplemented with 10% FBS. The cells were passage after achieving

approximately 80% confluence, and were used for experiments until the sixth to eighth









passage. Incubation of all cell lines described was at 37C under a humidified atmosphere

of 5% CO-,-95% air.

Immunohistochemistry. For immunofluorescence assays, cells were transferred to

22 x 22 mm sterilized Coming glass microscope cover slips, by placing the cover slips in

the wells of the Falcon six-well cluster trays, and plating the cells, which were then

allowed to reach 60 to 70% confluence. PAEC cells required pre-treatment of the cover

slips with 7-10 utg/ml fibronectin, as described above for culture dishes. Following three

5-min washes with phosphate buffered saline (PBS) to remove culture medium, cells

were fixed with a 4% paraformaldehyde solution for 20 min. To prepare the fixation

solution, 4% paraformaldehyde was added to 50% of the final volume of water and the

mixture was heated to 60-65C. Drop-wise addition of ION sodium hydroxide (usually 1-

2 drops) was required to completely dissolve the paraformaldehyde. After the solution

had cooled, 30% of the final volume of water and 20% of the final volume of 5X PBS

were added, and the solution was adjusted to a pH of 7.5 to 8.0. After incubation with the

cells, the fixative was removed with three 5-min washes in PBS, and any residual

paraformaldehyde was blocked with 50 mM glycine (in PBS) for 30 min. This

incubation was followed by three additional 5-min PBS washes. Paraformaldehyde

fixation was used for immunofluorescence experiments designed to examine plasma

membrane labeling. If cell permeabilization was desired in combination with

paraformaldehyde fixation, 0.1 to 0.2% Triton X-100 was added to the wells for the last

30 minutes of blocking. Alternatively, a -20C methanol incubation for 5-min was used

to fix cells for the purpose of intracellular staining. Following either fixation procedure,









cells were incubated in a solution of PBS containing 20% normal goat serum (NGS) and

3% bovine serum albumin (BSA), for 1-2 h in order to block non-specific antibody

binding. For pre-immune or immune labeling, cover slips were removed from wells and

inverted onto 50 p1 drops of pre-immune or primary antibody solution. The primary

antibodies were prepared in 20% NGS/PBS with 3% BSA (plus Triton-X when

appropriate), and for peptide competition assays, allowed to incubate with 50 jIg/ml of

corresponding peptide overnight at 4C. All incubations were at room temperature

(unless otherwise indicated), and antibody reactions were performed on parafilm in a dark

humid box. The dark humid box was prepared by placing PBS-saturated gauze across the

bottom of a small Tupperware container. Following a 2 h pre-immune or primary

antibody incubation, cells were returned to the wells and washed three times with PBS.

The secondary antibody, prepared in 20% NGS/PBS with 3% BSA (plus 0.1-0.2%

Triton-X when appropriate), was applied in a similar manner to the primary, and allowed

to incubate for 1 h before unbound molecules were removed with three 5 min PBS

washes. Following the last wash, cover slips were mounted onto glass slides with a drop

of Fluoromount-G, allowed to dry, and the edges of the cover slip sealed with fingernail

polish.

The secondary antibodies used in the immunofluorescence assays were either goat

anti-mouse or goat anti-rabbit IgG conjugated to either Texas Red or FITC fluoresceinn

isothiocyanate) fluorochromes (unless otherwise stated). Texas Red is excited when

exposed to a wavelength of 593 nm and emits a red fluorescence at a wavelength of 612

nm. FITC is excited at 494 nmn and emits a green fluorescence at 517 rnm. Because of the








different fluorescent properties, these two secondary antibodies can be used in double-

labeling experiments to show co-localization of two different proteins. However, the

proteins of interest must be detected using primary antibodies generated in two different

species. For example, double-labeling can be performed by incubating cells with a

primary antibody raised in mouse, and detected with a secondary goat anti-mouse Texas

Red antibody, and with another primary antibody raised in rabbit, detected using a goat

anti-rabbit FITC antibody.

Fluorescence light microscopy. Slides were initially viewed using a Nikon

Axiophot epifluorescence inverted microscope. A Leitz Planapo 63x, NA/1.4 oil

immersion lens and a modified Zeiss Axiomat inverted light microscope was used for

collecting three-dimensional light microscopy data sets (Agard, 1984; Agard and Sedat,

1983). The focal position, UV excitation shutter, and digital camera shutter of the

microscope were under computer control. The images generated were digitized directly

from the microscope image plane using a 14 bit, liquid nitrogen-cooled charge-coupled

device (CCD) digital camera (described in detail in Hiraoka et al., 1987; Agard et al.,

1989; Paddy et al., 1990). Three-dimensional data sets were collected as a series of

images separated by 0.5 mm along the horizontal optical sectioning axis (this value varies

with the depth of the cell). For double-labeling experiments consisting of different

fluorescence wavelengths, a complete 3-D data set at the first wavelength was collected,

the focus was returned to the starting focal position and the barrier filter was changed

using the computer control, then the second data set was collected (Hiraoka et al., 1991).

After data collection, each 3-D data set was corrected for stage and/or sample drift,

fluorescence photo-bleaching through the data set, and lamp intensity and/or shutter open








time variations. Following image correction, 3-D deconvolution corrected for the out-of-

focus contamination from each optical section. The images were displayed using an

integrated, multiple-windowed, mouse-driven display and Delta Vision software (Applied

Precision, Issaquah, WA).

Expression of exogenous transporter by transfection. The conditions for

transfection of the transporter cDNAs were optimized using human fibroblasts. The same

transfection protocol was used for Hela, PAEC, and HepG2 cells. Cells were plated onto

cover slips in 6-well trays 24 h before transfection. Optimal density for transfection was

within the range of 60-80% confluence for all cell lines used. Cells were transfected

using lipofectamine in Opti-MEM I reduced-serum medium according to the standard

protocol from Gibco Laboratories. For each transfection, 1 gg of a cDNA was added to

100 tl of Opti-MEM in a 17 x 100 mm polystyrene tube, and 4 gll of lipofectamine

reagent was mixed with 100 gl of Opti-MEM medium in a second polystyrene tube. The

solutions in each tube were combined and allowed to incubate for 30 min at room

temperature in order for cDNA-liposome complexes to form. During the incubation, cells

were washed two times quickly with PBS. For each transfection, 0.9 ml of serum- and

antibiotic-free MEM medium was mixed with the cDNA-liposome complexes and 1 ml

of the final transfection solution was applied to the cells. The cells were incubated with

the mixture for 3 h at 37C before washing two times with PBS and adding MEM

supplemented with antibiotics and 10% FBS. After 24 h, cells were fixed and labeled

according to the immunofluorescence protocol as described above. This lipofectamine

protocol above was compared to the liposome-mediated transfection protocols from either









Quiagen or Boehringer Mannheim and proved to be the least cytotoxic and provide the

greatest transfection efficiency (about 15-20%). To remove endotoxins from the cDNA

preps, cDNA in solution was mixed with 1% Triton X- 14, vortexed, and chilled on ice

for 5 min. The sample was heated for 5 min at 37C, and centrifuged at 14,000 x g for 5

min before recovering the aqueous solution to be used in the transfection protocols.

Preparation of transporter cDNA-Green Fluorescent Protein constructs. The

pEGFP vectors from Clontech (Palo Alto, CA.) encode the Green Fluorescent Protein

(GFP) variants that fluoresce 35 times more intensely than wild-type and have been

codon-optimized for maximal translation efficiency in mammalian cells (Cormack et al.,

1996). These vectors are available in all three reading frames and contain 20 unique

restriction sites in the multiple cloning region to facilitate subcloning. The pEGFP(N3)

Protein Fusion Vector from Clontech was used to fuse the EAAT1 cDNA to the N-

terminus of EGFP(N3). The EAAT1 cDNA in pCDNA3 was obtained from Dr. Jeffrey

Rothstein's laboratory. PCR primers were designed to the 5' terminus beginning at the

ATG translation start site (ATGACTAAAAGCAATGGAGAAGAGC) and the 3'

terminus ending at nucleotide 1680 (CATCTTGGTTTCACTGTCGATGG). Using 5 ng

of EAAT 1 cDNA and 100 pmol of each primer, the entire coding region minus the stop

codon was amplified with Taq polymerase according to the manufacturer's protocol

(InVitrogen, Carlsbad, CA). Amplification proceeded for 30 cycles of the following

conditions: denaturation at 94C for 1 min, annealing at 45C for 1 min, and extension at

72C for 1 min, with a final extension of 10 min. After the PCR product was obtained,

the 1680 base pair band was gel purified according to the protocol of Quiagen (Valencia,

CA) and cloned into the TA cloning vector, pCR 2.1 according to the manufacturer's








protocol (InVitrogen). A partial digest was performed using EcoRI restriction enzyme to

isolate the EAAT1 1680 base pair fragment. This fragment was then subcloned into the

pEGFP(N3) vector, at the EcoRI site, in order to place the EAAT1 cDNA before the

GFP. The 1680 base pair fragment was also cloned into the EcoRI site of the pEGFP(C3)

vector, which placed EAAT1 just before a stop codon at the C-terminus of the GFP.

A 2280 base pair fragment, including the entire coding sequence and stop codon,

of the CAT1 cDNA was cut out ofpCDNA3 with HindIII and EcoRI and subcloned

directly into the multiple cloning site ofpEGFP(C3) at the Hindll and EcoRI sites, thus

placing CAT1 at the C-terminal end of GFP.

Mutagenesis. Oligonucleotide-directed site-specific mutagenesis of the CAT1

cDNA was performed using the MORPH Mutagenesis kit from 5 Prime -- 3 Prime, Inc.

(Boulder, CO). CATMUT1 was constructed using the wild type GFP(C3)-CAT1 cDNA

as the template and the CAT 1.1 mutant oligonucleotide, GGT GTT GAG GGA GCG

GGA CAG GCG GCT CTC CTC CCG GCT GGA GTC GAC CAC CTT CCG GCG.

This mutagenesis reaction resulted in the substitution of serine for cysteine at residues 20

and 30. CATMUT2 was constructed using the GFP(C3)-CAT1 cDNA as the template

and the CAT1.2 mutant oligonucleotide, (GCA TCT GCT GGC CCA GCC CGA GCA

GGT TTT TGG AGGCCA TTG TGC TGA GCG AAT CTG C). This reaction resulted

in the substitution of alanine for glycine at position 2. CATMUT3 was generated using

CATMUT1 as the template and the mutant oligonucleotide, CAT1.3 (GCA TCT GCT

GGC CCA GCC CGA GCA GGT TTT TGC AGG CCA TTG TGC TGA GCG ATT

CTG C). This reaction resulted in the substitution of alanine for glycine at position 2,








and serine for cysteine at positions 3, 20. and 30. Each mutagenic oligonucleotide above

was designed with an internal restriction site for analyzing the success of the mutagenesis

reaction. A Sal I restriction site was engineered into the CAT 1.1 mutagenic

oligonucleotide and the CAT1.2 and CAT1.3 mutagenic oligonucleotides were

constructed with an internal Ava I site. Prior to beginning the mutagenesis, 2.5 P.g of the

oligonucleotide was 5' phosphorylated in a reaction using 5 p1 of 10X T4 polynucleotide

kinase buffer, 25 U of T4 polynucleotide kinase, and 10 mM ATP. This reaction was

allowed to proceed at 37C for 1 h before being terminated by heating to 65C for 10 min.

The annealing procedure involved mixing 0.03 pmol of the target cDNA (GFP-CAT1),

2.0 p.1 of 10X MORPH annealing buffer, and 100 ng of phosphorylated mutagenic

oligonucleotide and heating the solution to 100C for 5 min to denature the double-

stranded DNA. At this point, solutions were either placed at room temperature for 30

minutes, or placed in a beaker of 70C water that was allowed to cool slowly to room

temperature. One procedure worked better than the other for specific oligonucleotides

and the choice was determined experimentally. For the replacement strand reaction, 8 p.l

of3.75X MORPH synthesis buffer, 3 U T4 DNA polymerase, and 4 U T4 DNA ligase

were added directly to the annealing reaction and incubated for 2 hr at 37C, then for 15

min at 85C to terminate the reaction. A 1:10 dilution of Dpn I restriction enzyme was

prepared and 1 l1 was added to each mutagenesis reaction. The solution was incubated

for 30 min at 37C and then placed on ice for 5 min. The premise behind this digestion is

that Dpn I will specifically digest only double-stranded DNA in which both strands are

methylated, therefore, any double-stranded non-mutagenized target plasmid DNA will be








cleaved into small linear strands and will not be efficiently introduced and propagated

during the bacterial transformation. Following the digestion, the entire mutagenesis

reaction was added to 200 utl ofE. coli MORPH mutS cells and incubated on ice for 20

min. The cells were heat-shocked at 42C for 2 min and spread on LB plates containing

30 Htg/ml kanamycin for selection of the appropriate target plasmid. The mutS strain of

E. coli is used because it is deficient in DNA repair strand selection, therefore, it

randomly repairs either the mutant strand or the original template pair. This way, there is

a 50% chance that the mutant strand will be selected as correct and that sequence will be

propagated further. After selecting colonies and growing the bacteria in LB plus 30

jig/ml kanamycin, the plasmid DNA was isolated from several colonies with Quiagen

minipreps and digested using the enzyme specific to the engineered sequences to confirm

they contained the mutant strand (Sal I for CATMUT1 and Ava I for CATMUT2 and

CATMUT3). Large scale plasmid prep kits (manufactured by Quiagen) were used to

prepare the final mutant cDNAs. Each of the mutants was subjected to a series of

restriction digests to confirm the amino acid substitution(s) and to check for appropriate

sizes of the vectors and inserts.

Data analysis. Much of the data generated in the proposed experiments were

qualitative and required visual analysis. Each immunofluorescence experiment was

performed a minimum of three times to check for consistency among cell populations

plated on different days. In addition, normal and LPI cells of the same passage number

were assayed in parallel using the same reagents and antibody solutions. LPI cells from

several different patients were used for this study. Cells from the same patient were used






27

to check for reproducibility, then cells from a different patient were used to confirm that

the results are not unique to a single individual with the disease.














CHAPTER 3
DISTRIBUTION OF THE GLUTAMATE TRANSPORTERS

Introduction

Five plasma membrane proteins have been cloned that belong to the family of

transporters responsible for the Na--dependent uptake of glutamate and aspartate in a

variety of tissues (see Chapter 1 for overview). These transporters correspond to the

high-affinity, Na--dependent System XAG" activity previously described in human

fibroblasts (Dall'Asta et al., 1983) and cultured liver cells (Makowske and Christensen,

1982). The glutamate transporters share approximately 50% amino acid identity, consist

of 6 or more transmembrane spanning domains, and contain two experimentally

identified N-linked glycosylation sites on the second extracellular loop (Conradt et al.,

1995). Glutamate uptake by these transport proteins is electrogenic. and coupled to the

co-transport of three Na" ions and one H', as well as the counter-transport of one K' ion

(Zerangue and Kavanaugh, 1996).

Although similar in structure and substrate specificity, the glutamate/aspartate

transporters are differentially expressed and regulated. In the brain, EAAT1 (GLAST1)

and EAAT2 (GLT1) are localized to astroglia, whereas EAAT3 (EAAC1) is specific to

neurons. EAAT4 also has been localized to neurons and EAAT5 is specific to retinal

tissue. The family members also differ with respect to their transport properties. When

ooyctes were injected with EAAT5 cRNA, radiolabeled glutamate uptake was increased








by 2- to 10-fold over that of the control uninjected oocytes (Arriza et al., 1997).

However, this was significantly less than the transport observed with EAAT1, EAAT2.

and EAAT3-expressing oocytes. which were reported to transport 50-fold over the level

of uninjected oocytes (Klockner et al.. 1993; Kanai et al., 1995). Whereas

electrophysiological studies have shown that both EAAT4 and EAAT5 transport

glutamate poorly, they possess much stronger chloride channel properties only weakly

present in the other glutamate transporters (Fairman et al., 1995; Arriza et al., 1997).

Several laboratories have documented the involvement of glutamate in cell migration and

differentiation (Pearce et al., 1987; Mattson et al., 1988), as well as neuronal and

astroglial proliferation. Osnat Bar-Peled et al. have shown that each of the glutamate

transporters has a specific and unique distribution during brain development (Osnat Bar-

Peled et al., 1997), suggesting that the transporters play multiple functional roles during

brain maturation.

Three amino acid residues in the C-terminal sequence, the region of greatest

homology, have been identified as essential for glutamate transport activity (Conradt and

Stoffel, 1995; Pines et al., 1995). Using site-directed mutagenesis. Pines and coworkers

determined that aspartate 398, glutamate 404, and aspartate 470 are critical for EAAT2

activity, and glutamate 404 may contribute to substrate specificity (Pines et al., 1995).

Both aspartate 398 and 470 appear to be involved in transporter activity, rather than

stability or trafficking, and even the conservative replacement of glutamate abolishes

transport activity. Zhang et al. showed, using site-directed mutagenesis, that histidine

326 is required for glutamate transport by EAAT2, and probably contributes to the proton

translocation mechanism that accompanies the Na'- and KI-coupled transport activity








(Zhang et al.. 1994). Conradt and Stoffel performed similar mutagenesis studies using

the EAAT1 cDNA (Conradt and Stoffel, 1995). When they mutated the conserved

arginine 122, arginine 280, arginine 479, and tyrosine 405, they lost glutamate transport

activity with the tyrosine 405 and arginine 479 mutants. The arginine 122 and arginine

280 mutants appeared to increase the Km of EAAT1 for aspartate. but had no effect on the

intrinsic properties or kinetics of glutamate transport. They proposed from their studies

that the hydroxyl group on tyrosine 405 and the positive charge on arginine 479 may

contribute to the binding of the acidic glutamate substrate.

Although most of the research involving the glutamate transporters has been

confined to the brain, L-glutamate is crucial to several biochemical pathways of

peripheral tissues as well (i.e., ammonia detoxification and gluconeogenesis). Several

laboratories have independently shown, by mRNA and protein analysis, that tissues other

than the brain express one or more of the transporter isoforms. EAAT3 is the most

ubiquitous of the glutamate transporters and is detected in kidney, small intestine, liver,

heart, lung, skeletal muscle, and placenta (Kanai and Hediger, 1992; Matthews et al.,

1998). Glutamate and asparate are almost completely reabsorbed from the glomerular

filtrate by the abundantly expressed EAAT3 transporter in the renal tubules (Silbemrnagl,

1983; Shayakul et al., 1998). EAAT1 is expressed in heart, lung, skeletal muscle, retinal

glia, and placenta (Kawakami et al., 1994; Arriza et al., 1994), and both EAAT2 and

EAAT4 have also been detected in placental tissue (Matthews et al., 1998). Preliminary

data from our laboratory (Tessmann, unpublished results) also suggest that EAAT1,

EAAT2, and EAAT3 are expressed in human fibroblasts.








L-Glutamate and L-aspartate are important nutritional substances that contribute

to a variety of biochemical pathways in the brain and peripheral tissues. Specialized

glutamatergic neurons in the brain produce and store glutamate. the major excitatory,

neurotransmitter in the mammalian central nervous system, until it is released into the

synaptic cleft in response to different stimuli. Intracellular concentrations of glutamate in

the brain reach approximately 10 mM, with the highest concentrations at nerve terminals

(Shupliakov et al., 1992; Storm-Mathisen et al., 1992). Extracellular levels of glutamate

are carefully maintained below 3 [tM, except during neurotransmission of a signal when

concentrations may reach between 1-2 mM (Nicholls, 1993: Clements et al.. 1992).

Members of the EAAT family of transporters, primarily the glial-specific EAAT1 and

EAAT2 (Rothstein et al., 1996). are responsible for the high-affinity Na--dependent

transport of glutamate out of the synaptic cleft against a thousand-fold concentration

gradient. This carefully regulated transport activity is crucial for maintaining glutamate

concentrations below the level that is toxic to neurons.

Glutamatergic transmission is believed to contribute to normal brain activities

such as learning and memory (Bliss et al., 1993), however, elevated levels of extracellular

glutamate are neurotoxic and can lead to several neuro-degenerative diseases such as

Amyotrophic Lateral Sclerosis (ALS), Huntington's disease, and probably Alzheimer's.

In 1992. Rothstein et al. showed that brain and spinal cord samples taken from the

autopsies of ALS patients revealed a reduced level of glutamate uptake (Rothstein et al.,

1992). More recently it was determined that the decrease in glutamate uptake in some

cases of sporadic ALS is due to the selective loss of the astroglial EAAT2 glutamate








transporter (Rothstein et al.. 1995). It has recently been proposed that the down-

regulation of EAAT2 results from a defect in mRNA processing. Lin et al. has shown

that due to defective mRNA splicing events such as intron-retention or exon-skipping.

multiple abnormal EAAT2 mRNA species are produced in the affected areas of the brain

in ALS patients (Lin et al., 1998). In vitro expression studies suggested that the protein

products of the aberrant mRNAs had decreased transport activity because they were

degraded rapidly, or perhaps, had a dominant-negative effect on the normal EAAT2

protein. The abnormal mRNA species were not present in regions of the brain that were

unaffected in ALS patients, or in the non-neurologic disease controls.

Decreased glutamate transporter activity in the frontal, parietal, and temporal

cortex has been implicated in the neurodegeneration that occurs in Alzheimer disease

(Scott et al., 1995: Cowbum et al., 1988). Like ALS, mRNA levels ofEAATI. EAAT2.

and EAAT3 were normal in the frontal cortex, however, immunoblot analysis detected

about 30% less EAAT2 protein (Li et al., 1997). On the other hand, schizophrenia and

other psychoses are thought to result partially from glutamatergic hypofunction, a

condition that occurs following excessive glutamate uptake (Carlsson and Carlsson.

1990). Therefore, the mechanism by which glutamate is cleared from the synaptic cleft

must be tightly regulated in order to prevent neuronal damage or malfunction.

The roles that the individual transporters play in normal synaptic clearance and

neurotoxicity are unclear because subtype-specific inhibitors are not available. This has

led a number of laboratories to study the functions of these specific transporters using

antisense oligonucleotides or knockout mice. Results from antisense studies indicated

that a loss of EAAT1 and EAAT2 resulted in elevated extracellular glutamate








concentrations as well as neurodegeneration and progressive paralysis (Rothstein et al..

1996). Antisense oligonucleotides to EAAT1. EAAT2. and EAAT3 were administered

intraventricularly to male Sprague-Dawley rats for 7-10 days. Within 3 days. the animals

that were treated with the EAAT1 and EAAT2 antisense oligonucleotides began to show

evidence of progressive motor degeneration, and by day 8. their hindlegs were paralyzed.

When extracellular levels of glutamate were measured with microdialysis probes in the

ipsilateral striatum of treated rats, extracellular glutamate concentrations were elevated by

32-fold in EAAT2 antisense rats, and by 13-fold in EAAT1 antisense rats. Loss of

EAAT3 did not elevate extracellular glutamate levels in rats treated with EAAT3

antisense oligonucleotides. These animals did, however, experience epileptic seizures and

slightly impaired motor skills in some cases. These data support the findings that

EAAT1 and EAAT2 play a crucial role in synaptic clearance of glutamate, however, the

role of EAAT3 in preventing neurological damage seems to be unclear due to conflicting

data.

Evidence suggests that EAAT2 is responsible for the greatest amount of cerebral

glutamate transport (Robinson, 1991). When EAAT2 was knocked out by homologous

recombination, levels of residual glutamate increased in the brain and the mice suffered

lethal spontaneous seizures (Tanaka, 1997). Both the antisense and knockout studies are

consistent with the data documenting the loss of EAAT2 as being the major cause of the

motor neuron degeneration that plagues patients with ALS (Rothstein, 1995).

Conversely, when EAAT3 (EAAC 1) knockout mice were produced, no

neurodegeneration was observed (Peghini and Stoffel, 1997). Instead, the mice

developed dicarboxylic aminoacidurea. analogous to an inborn error of glutamate and








aspartate transport across epithelial cells of the kidney and intestine. In some cases.

mental retardation or neurological abnormalities are also symptoms of the disease.

however, there was no evidence of neurological damage in the EAAT3 knockout mice.

The dicarboxylic aminoacidurea acquired by the knockout mice is explained by the fact

that EAAT3 is strongly expressed in the kidney and is responsible for tubular

reabsorption of glutamate and aspartate from the glomerular filtrate.

Based on their accepted function of Na'-dependent glutamate/aspartate transport

and the membrane spanning structure of these amino acid transporters, we predicted that

the EAAT transporters would be primarily localized to the plasma membrane, with some

intracellular pools involved in biosynthesis or recycling. However, this could not be

assumed based on the recent detection of several plasma membrane proteins in the

nuclear membrane and matrix. P-glycoprotein, a 170 kDa protein with 12 membrane

spanning domains, has been implicated in conferring multi-drug resistance in cancer cells

(Juliano and Ling, 1976). It accomplishes this by actively exporting a wide-range of

chemotherapeutic agents out of the cell in an ATP-dependent mechanism. Recently. P-

glycoprotein was reported to reside in the nuclear membrane and matrix as well as on the

plasma membrane (Baldini, 1995). Also, a number of laboratories have reported the

presence of different growth factor receptors associated with the nucleus (Stachowiak et

al., 1996; Podlecki et al., 1987; Rakowicz-Szulczynska et al.. 1989). Traditionally, these

receptors were believed to reside on the plasma membrane and transmit a signal from the

extracellular environment to the cytosol upon ligand binding. However, the fibroblast,

insulin, and epidermal growth factor receptors are all integral membrane proteins that

have been localized to the nucleus.








In a study of the activity and localization of the glutamate transporters in day 14

and day 20 rat placenta, Matthews et al. detected EAAT1 in the nuclei of the maternal

decidua and placental trophoblast (Matthews et al.. 1998). In other cell types of the

placenta, however. EAAT1 was detected on the plasma membrane and in one or more

intracellular vesicle populations. EAAT1. EAAT2, and EAAT3 mRNA and protein

expression were increased in the day 20 placenta, and the expression patterns were cell-

type specific for each isoform. Neither EAAT2 nor EAAT3 was observed in the nuclei of

any of the cell types examined by immunohistochemistry. Although EAAT4 mRNA was

identified in both day 14 and day 20 placenta, EAAT4 protein was not detectable in either

sample by immunoblot or immunohistochemistry.

Over the past two decades, extensive research has been performed to determine

the ionic requirements, substrate specificity and kinetic parameters of the glutamate

transporters, however, there is almost no information regarding intracellular pools.

trafficking through compartments, or protein arrangement on the plasma membrane.

One of my intentions for this project was to answer some basic questions regarding the

cellular and subcellular distribution of the different isoforms. Using sequence-specific

antibodies generated against the individual EAAT transporters, I could explore the

expression of the isoforms in cell lines other than the brain. In addition, I could

determine if the different EAAT family members display a similar or distinct staining

pattern in regard to abundance and distribution of protein. This chapter describes the

results obtained from staining human fibroblasts with antibodies against the glutamate

transporters EAAT1 and EAAT3. More extensive work involving co-localization

experiments in several cell lines was performed using EAAT1 antibodies from several








independent sources. The EAAT1 cDNA was used to create a fusion protein with the

green fluorescent protein (GFP) and distribution of the transporter was examined

following transient expression of this EAAT1-GFP fusion protein. These experiments

were performed in combination with immunofluorescence and immunoblotting using four

different antibodies specific to EAAT1.

Methods

Glutamate transporter antibody production. The polyclonal EAAT3 glutamate

transporter antibody was raised in rabbit against an EAAT3-maltose binding protein that

was constructed by Dr. Marc Malandro using the C-terminal 120 amino acids of the rat

EAAT3 (EAAC 1) sequence. The glutamate transporter antibody designated EAAT1-R

was obtained from Dr. Jeffrey Rothstein at Johns Hopkins University. EAAT I -R is a

polyclonal antibody that was raised in rabbit against the amino acid residues 3-17

(KSNGEEPRMGSRMGR) at the N-terminus of the human EAAT1 protein (Ginsberg et

al., 1995). The EAATI glutamate transporter antibody designated EAAT1-S was

obtained from Dr. Wilhelm Stoffel at the University of Cologne (Cologne, Germany).

This polyclonal antibody was generated in rabbit against a peptide consisting of amino

acid residues 24-40 (KRTLLAKKKVQNITKED) at the N-terminus of the rat EAAT1

sequence (Wahle and Stoffel, 1996). The EAAT1-C polyclonal antibody was generated

in guinea pig, by Chemicon International, Inc. (Temecula, CA), against the rat C-terminal

peptide, QLIAQDNEPEKPVADSETKM (Storck et al., 1992). The polyclonal EAAT1-

D antibody was purchased from c-Diagnostics International (San Antonio, TX). This

antibody was raised in rabbit against a peptide (amino acids 504-518,








NRDVEMGNSVIEENE) from the C-terminus of the rat EAAT1 sequence (Rothstein et

al.. 1995).

Cell fractionation. Media was aspirated from human fibroblasts grown to 80-90%

confluence in 150-mm dishes. Cells were washed three times with 10 ml ice cold PBS or

SEB (85.6 g/L sucrose. 0.76 g/L EGTA, 2.38 g/L Hepes, pH 7.5). then scraped and

collected in a plastic 50 ml centrifuge tube. A total of 15 ml of SEB with 0.5 mM

phenylmethyl sulfonyl fluoride (PMSF) and 1 pl/ml protease inhibitors (1 pLg/ml antipain

and leupeptin, and 100 KIU/ml aprotinin) per 150-mm dish was used to scrape the cells.

Using a refrigerated table-top centrifuge, cells were spun at 300 x g for 5-10 minutes,

supernatant discarded, and cell pellets resuspended in 15 ml of SEB + PMSF and protease

inhibitors. Cell suspension was poured into an ice-cold nitrogen bomb (Parr Instrument

Company. Moline, IL) and allowed to equilibrate at 200 psi for 10 min before lysis by

rapid release. Resultant homogenate suspension was collected in a centrifuge tube and

checked for unbroken cells under a light microscope. Homogenate was centrifuged at

300 x g for 10 min as described above, and the pellet was saved as the nuclear fraction.

The 300 x g supernatant was spun in an ultracentrifuge at 15,000 x g (14,500 rpm in a

Beckman 60TI rotor) for 30 min and the pellet saved as a crude plasma membrane-

enriched fraction. The supernatant from the 15,000 x g spin was centrifuged in the

ultracentrifuge at 100.000 x g (37.500 rpm in a Beckman 60TI rotor) for 60 min and the

pellet was saved as the total intracellular membrane fraction.

SDS-PAGE and immunoblot analysis. Samples from the cell fractionation

procedure were initially dissolved in 0.2 N NaOH/0.2% SDS, and then further diluted to








0.5-1.0 ig/tl in sample dilution buffer and 10-20 tg protein per lane were subjected to

one-dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE)

(Laemmli, 1970; Chiles et al., 1987). The proteins were transferred at 299 mAmps for 18

h to a nitrocellulose membrane (Chiles et al., 1987). The nitrocellulose membrane was

blocked with 5% non-fat dry milk (NFDM) at room temperature for 1.5 h. rinsed with

TBS/TWEEN (30 mM Tris base, 150 mM NaCI, 0.1% Tween 20, pH 7.6), then incubated

with primary antibody (summarized below) prepared in TBS/TWEEN for 1-2 h at room

temperature. Two quick rinses, one 15 min rinse, and two 5 min rinses with

TBS/TWEEN were followed by incubation of the nitrocellulose in the secondary

antibody, prepared in TBS/TWEEN, for 1-2 h at room temperature. The EAAT1-R

antibody was used at 1:200-1:1000 dilutions, the EAAT1-S antibody was used at a 1:50

dilution, and both were visualized using a 1:5000 dilution of either a donkey or goat anti-

rabbit IgG conjugated to horseradish peroxidase. The EAAT1-C antibody was used at a

dilution of 1:100 and detected with a 1:10,000 dilution of goat anti-guinea pig IgG

conjugated to horseradish peroxidase. After washing the nitrocellulose blot six times for

5 min each with TBS/TWEEN, the blot was incubated in 6 ml of a 1:1 mixture of the

Enhanced Chemiluminescence (ECL) reagents (Pierce, Rockford, IL) for 1 min, drained.

wrapped in plastic, and immediately exposed to film.

Antibodies for immunoblotting and immunofluorescence. The antibodies used for

immunoblotting (described above) and immunofluorescence (described in Chapter 2) are

summarized in Table 3-1. Unless otherwise stated, EAAT1 and EAAT3 antibodies were

detected during immunofluorescence assays using a 1:200 dilution of goat anti-rabbit IgG








conjugated to fluorescein isothiocyanate (FITC). For double-labeling experiments.

monoclonal antibodies against the organelle-specific proteins were detected using a 1:200

dilution of goat anti-mouse IgG linked to Texas Red. The results of all

immunofluorescence experiments in this chapter were analyzed using deconvolution

microscopy (described in the Methods Chapter).

Table 3-1
Antibodies against Glutamate Transporters
Name Host Source Dilutions
IF* IB**
EAAT3 rabbit Dr. Michael Kilberg, 1:200 --
Univ. of Florida
EAATI -R rabbit Dr. Jeffrey Rothstein, 1:50 1:200-
Johns Hopkins 1:1000
EAAT1-S rabbit Dr. Wilhelm Stoffel, 1:50 1:50
Univ. of Cologne, Germany
EAAT 1-C guinea pig Chemicon International. 1:1000 1:100
Temecula, CA
EAAT1-D rabbit x-Diagnostics, 1:50 --
San Antonio, TX
KDEL mouse Dr. David Vaux, 1:100 --
EMBL
Transferrin receptor mouse Zymed, 1:5 --
San Francisco, CA
414 mouse Dr.JohnAris, 1:10 --
Univ. of Florida
D77 mouse Dr. John Aris, 1:50 --
Univ. of Florida
*IF = immunofluorescence assay
**IB = immunoblot




Results


Localization of endogenous EAAT3 glutamate transporter in human fibroblasts by

immunofluorescence. Immunofluorescence assays were performed on human fibroblasts

using antibodies generated against the glutamate transporters, EAAT1 and EAAT3.








Preliminary experiments with antibodies specific for EAAT2 and EAAT4 revealed no

detectable staining (data not shown). These experiments were intended to provide some

information concerning the similarities and differences in cellular distribution of amino

acid transporters belonging to the same gene family, and perhaps gain insight into why

multiple transporters with nearly identical kinetics are expressed in the same cell. For all

extracellular labeling experiments, fibroblasts were fixed with a 4% paraformaldehyde

solution and stained according to the immunofluorescence protocol described in the

Methods Chapter.

A 1:200 dilution of the anti-EAAT3 antibody stained the cells in clusters rather

than diffusely labeling the entire cell surface (Figure 3-1 A). This pattern of transporter

"patching" was similar to the pattern observed using the CAT1 arginine transporter

antibody (described in Chapter 5. Figure 5-1 A and B). Labeling by the anti-EAAT3

transporter antibody was completely inhibited by preadsorption of the antibody with 50

ptg/ml of the corresponding peptide antigen for 12 h at 4C (Figure 3-1 B). The

intracellular distribution of the EAAT3 glutamate transporter was examined in human

fibroblasts following either -20C MeOH fixation, or 4% paraformaldehyde fixation in

combination with 0.1% Triton X-100 membrane permeabilization. Under both

conditions, the EAAT3 antibody labeled small vesicles throughout the cytoplasm (Figure

3-2 A). Double-labeling MeOH-fixed fibroblasts with antibodies against EAAT3 (1:200

dilution) and KDEL (1:100 dilution), a common epitope of resident proteins of the

endoplasmic reticulum (ER), showed very little co-localization (Figure 3-2 B). Also,

only a small amount of co-localization was detected when MeOH-fixed fibroblasts were








double-labeled with a 1:200 dilution of the EAAT3 antibody and a 1:5 dilution of the

transferring receptor (Figure 3-2 C). indicating that very little of the EAAT3 glutamate

transporter is involved in recycling under standard culture conditions.

Localization of endogenous EAAT1 glutamate transporter in human fibroblasts by

immunofluorescence. Initial experiments were performed using the EAAT 1-R antibody

generated in Dr. Jeffrey Rothstein's laboratory against an N-terminal peptide sequence.

This antibody will be referred to as EAAT1-R (see Table 3-1 for details on EAAT1

antibodies). Unlike the "patches" observed on human fibroblasts with the anti-EAAT3

antibody, the anti-EAAT 1 -R antibody did not detect any protein on the surface of

paraformaldehyde-fixed fibroblasts even at a 1:25 antibody dilution. When fibroblasts

were fixed with -20C MeOH and stained for intracellular EAAT1, using a 1:50 dilution

of the anti-EAATI -R antibody, the nuclear membrane and nuclear matrix were the

primary structures detected (Figure 3-3 A). Small vesicles throughout the cytoplasm,

resembling the vesicles labeled with the EAAT3 antibody, were also apparent. Staining

of the anti-EAAT I -R transporter antibody was completely inhibited by preadsorption of

the antibody with 50 tg/ml of the corresponding peptide antigen for 12 h at 4C (Figure

3-3 B). To confirm that the EAAT1-R antibody was staining the nuclear membrane,

MeOH-fixed fibroblasts were double-labeled with antibodies against EAAT1 -R (1:50

dilution) and 414 (1:10 dilution) (Figure 3-3 C). The latter is an antibody generated

against an epitope shared by several proteins of the nuclear pore complex (Davis and

Blobel, 1986). There was significant co-localization of the EAAT1-R and 414

antibodies, strongly suggesting that EAAT1 -R was labeling the nuclear membrane. On








the other hand, very little co-localization was detected when EAAT I -R and KDEL

antibodies were used (data not shown), indicating that the fluorescence was not likely due

to staining of ER components surrounding the nucleus. A D77 antibody, generated

against the yeast nucleolar protein, fibrillarin (Noplp), was used at a dilution of 1:50 in

double-labeling experiments with EAAT1 -R antibody (1:50 dilution) to determine if the

nuclear staining was nucleolar (Aris and Blobel, 1988). The D77 antibody detected three

or four nucleoli in each cell, however, no co-localization with EAAT1-R was observed

(Figure 3-3 D).

The nuclear membrane staining of human fibroblasts with an antibody against a

known plasma membrane amino acid transporter (EAAT1) was unexpected. In an effort

to confirm the nuclear localization of the EAAT1-R antibody, a second EAAT1 antibody

was obtained from Dr. Wilhelm Stoffel (University of Cologne, Germany) and will be

referred to as EAAT1-S (Table 3-1). Although EAAT1-S was also generated against an

N-terminal peptide. there was no overlap with the amino acid sequence that EAATI -R

was raised against. Therefore, EAAT1-S provided an additional antibody in case the

EAAT 1-R antibody was cross-reacting to an unknown protein with homology to the

peptide used to generate EAAT1-R. When MeOH-fixed human fibroblasts were

incubated with a 1:50 dilution of the anti-EAAT1-S antibody, no nuclear membrane or

matrix staining was observed (Figure 3-4 A). Only small vesicles throughout the

cytoplasm were detected.

With the apparent contradicting results obtained from immunofluorescence

experiments using the EAAT1-R and EAAT1-S antibodies, two additional EAAT1








antibodies were purchased from Chemicon (EAAT1 -C) and a-Diagnostics (EAAT1 -D).

Both antibodies were incubated with MeOH-fixed human fibroblasts according to the

same procedures used for the EAAT1-R and EAAT1-S antibodies. Nuclear staining was

observed using a 1:1000 dilution of the EAAT1-C antibody (Figure 3-4 B). detected with

a 1:400 dilution of goat anti-guinea pig IgG linked to Cy3, but that labeling associated

with the nuclear membrane was fainter than with the EAAT1-R, and there also was

significant staining of a cellular vesicle population. Thus, the pattern of fluorescence

generated by the EAAT1-C antibody appeared to be a mixture of the results obtained

from the EAAT1-R and EAAT1-S antibodies. No specific or background staining was

detected when human fibroblasts were incubated with EAAT1-C primary or anti-guinea

pig secondary antibodies alone. However, attempts to enhance the EAAT1-C nuclear-

specific staining by centrifuging (5 min at 13.000 x g) or filtering the EAAT1 -C antibody

through a 2 um pore syringe prior to use failed. A 1:50 dilution of the EAAT1 -D

antibody resulted in a uniform, diffuse cytosolic labeling with no specific staining of the

nuclear membrane, cytoplasmic vesicles, or other cellular structures (data not shown).

Localization of the EAAT1 glutamate transporter in other cell types. To

determine whether or not the nuclear localization of EAATI was unique to fibroblasts,

several other cell types were labeled with both the EAAT1-R and EAAT1-S antibodies.

Hela, HepG2 (human hepatoma), and pulmonary artery endothelial cells (PAEC) were

each fixed with -20C MeOH and stained with 1:50 dilutions of both EAAT1-R and

EAAT1-S according to the immunofluorescence assay described in the Methods Chapter.

The EAAT1-R antibody labeled the nuclei of both Hela (Figure 3-5 A) and HepG2 (data








not shown), however, no nuclear staining was observed in the PAEC (Figure 3-9 A).

Instead, the EAAT1-R antibody stained small vesicles throughout the cytosol of the

PAEC, with slightly more fluorescence concentrated in the perinuclear region. As was

the case for the human fibroblasts, no nuclear staining was detected with the EAAT1-S

antibody in any of the cell lines. In Hela. HepG2. and PAEC, the EAAT1-S antibody

detected only small vesicles of unknown origin in the cytosol. Figure 3-5 (A and B)

shows representative Hela cells stained with the EAAT1-R and EAAT1-S antibodies. _

Identification of EAAT 1 glutamate transporter by immunoblot analysis. Total

intracellular membrane (100,000 x g), crude plasma membrane-enriched (15,000 x g),

and nuclear fractions (300 x g) from human fibroblasts, HepG2, and PAEC were isolated,

subjected to SDS-PAGE, and transferred to nitrocellulose membranes for

immunoblotting. The primary antibodies were detected using 1:2,500 to 1:10.000

dilutions of goat anti-rabbit IgG (EAAT1 -R and EAAT1 -S) or goat anti-guinea pig IgG

(EAAT1-C) conjugated to horseradish peroxidase (described in the Methods Section of

this chapter). When the 300 x g nuclear fraction from human fibroblasts was incubated

with a 1:1000 dilution of EAAT1-R (Figure 3-6), a strong band was detected at

approximately 70-75 kDa, corresponding to the molecular mass of the monomeric

EAAT1 protein (Tessmann and Kilberg, unpublished data). Less intense bands were

observed at higher molecular masses, including a light band at approximately 180 kDa,

which corresponds to the molecular mass of a putative EAAT1 trimer. EAAT1 was also

detected in the 100,000 x g membrane fraction, however, the protein was primarily

detected at approximately 180 kDa, which is probably the trimeric form. Our laboratory

observes that the more manipulation of the sample, such as the additional centrifugation








steps required to obtain the 100,000 x g pellet, the greater the shift from monomer to

dimer/trimer forms of the transporter. Also, detection of the higher molecular mass

species of EAAT I is consistent with published data from other laboratories documenting

the formation of glutamate transporter homomultimers (Haugeto et al.. 1996).

When HepG2 cells were subjected to cell fractionation, SDS-PAGE, and

immunoblotting (see Methods section above), a 1:200 dilution of EAATI-R was detected

in the 300 x g fraction in both the monomeric (approximately 70 kDa) and higher

molecular mass form (approximately 180 kDa) (Pappas and Kilberg, unpublished data).

EAAT1 -R immunoreactivity was detected in the 180 kDa form in the 15,000 x g and

100,000 x g fractions, as well as two smaller species (approximately 70 and 75 kDa) in

the 100,000 x g fractions (data not shown).

Cell fractionation, SDS-PAGE, and immunoblotting were also performed using

PAEC according to the protocols in the Methods Section of this chapter. To determine if

the EAAT1 immunoreactivity in the nuclear (300 x g) fraction was a result of

contamination by other membranes or unbroken cells, a 1:100 dilution of A-lA5 (I -

integrin) antibody was incubated with each of the fractions (data not shown). The A-1 A5

antibody detects one or more P-integrin species (depending on the cell type) that migrate

at 210, 165, and 130 kDa (Hemler et al., 1984), and are specific for the plasma

membrane. I have confirmed the plasma membrane specificity of the antibody by SDS-

PAGE and immunoblot analysis (data not shown). Bands of approximately 165 kDa

were detected by the 13-integrin antibody in the 300 x g fraction as well as the 15,000 x g

fraction of PAEC. Very little protein was observed in the 100,000 x g fraction. Although








the majority of the P3-integrin labeling appeared in the plasma membrane-enriched

fraction, the band in the nuclear fraction suggests that the 300 x g fraction is

contaminated with plasma membrane, or more likely unbroken cells. When a 1:100

dilution of EAAT1 -C was incubated with PAEC fractions, a band at approximately 75

kDa was detected in all three fractions in almost equal amounts (data not shown).

Transfection of EAATl-GFP constructs in human fibroblasts. Two fusion

constructs between EAAT1 and green fluorescent protein (GFP) were generated to

compare the localization of the expressed transporter with the endogenous EAAT 1

immunofluorescence experiments. The EAAT1-GFP(N3) fusion protein was constructed

with the GFP tag at the C-terminal end of EAAT1, and the GFP(C3)-EAAT1 was

constructed with GFP at the N-terminal end. The GFP tag was attached to either the N-

or C-terminus of EAAT 1 to ensure that the location of the GFP did not interfere with

normal transporter trafficking (see Methods Chapter for details). Both of these constructs

were expressed in human fibroblasts using the lipofectamine transfection procedure

described in the Methods Chapter. After 24 hours of expression, cells were fixed with

-20C MeOH, mounted on glass slides with Fluoromount-G, and visualized using an

FITC filter on the deconvolution microscope. Human fibroblasts transfected with the

GFP vector only (Figure 3-7 A) showed diffuse fluorescence throughout the entire cell,

whereas cells transfected with either EAAT1 -GFP(N3) (Figure 3-7 B) or GFP(C3)-

EAAT1 (Figure 3-7 C) showed distinct vesicles in the cytoplasm as well as staining of

the plasma membrane. After analysis of three separate experiments, it was concluded that

there was no difference in the fluorescent patterns generated by the two different GFP








fusion proteins. Although some perinuclear fluorescence was observed, probably Golgi

localization, the pattern was clearly distinct from the nuclear membrane labeling observed

with the EAAT1-R antibody.

Co-localization of EAAT1-GFP(N3) and the endogenous EAAT1 in human

fibroblasts. The fluorescence pattern observed in the EAAT1 -GFP(N3)-transfected

fibroblasts was consistent with the pattern of staining by the EAAT1 -S antibody, that is

fluorescence was associated with a population of small vesicles scattered throughout the

cytoplasm. However, to test the specificity of both the EAAT1-S and EAAT1-R

antibodies for the exogenous transporter, human fibroblasts were transfected with the

EAAT1-GFP(N3) fusion protein, fixed with -20C MeOH, and stained with either the

EAAT1-R (Figure 3-8 A) or EAAT1-S (Figure 3-8 B) antibody. The EAAT1-S and

EAAT1-R primary antibodies were detected using a 1:200 dilution of goat anti-rabbit IgG

conjugated to Texas Red. A significant amount of co-localization was observed with the

EAAT1-GFP(N3) fusion protein and EAAT1 -S antibody, however, little overlap was

detected when the EAAT1-R antibody was used. In the fibroblasts stained with EAAT1-

S there were separate pools of Texas Red-labeled vesicles that did not overlap with

EAAT1 -GFP(N3). This could be due to the EAAT1 -S antibody recognizing endogenous

EAAT1 which, for unknown reasons, was localized in a compartment lacking the

expressed fusion protein. Conversely, there was also a population of vesicles containing

EAAT1-GFP(N3) that were not stained with EAAT1-R. This latter result can be

explained by two vesicle populations or by the fact that the antibody-antigen binding

efficiency is less than 100% for the immunofluorescence assay. These results strongly








suggest that the EAAT1-S. but not the EAAT1-R. antibody recognizes the expressed

EAAT1-GFP(N3) protein in the human fibroblasts.

Co-localization of EAAT1-GFP(N3) with the EAAT1 endogenous in PAEC. As

an extension of the experiments in the last section. PAEC were transfected with the

EAAT1 -GFP(N3). fixed with -20C MeOH, and stained with either the EAAT1 -R (Figure

3-9 A) or EAAT1-S (Figure 3-9 B) antibody. Primary antibodies were detected using a

1:200 dilution of goat anti-rabbit IgG linked to Texas Red. PAEC were selected because

they were the only cell type tested that did not show nuclear localization of EAAT1 in the

initial experiments with the EAAT1-R antibody. As observed with the human

fibroblasts, there was significant co-localization between the EAAT 1-GFP(N3) fusion

protein and the EAAT1-S antibody in PAEC. Although the fluorescent patterns of the

EAAT1-R antibody and EAAT1-GFP(N3) were both punctate and cytoplasmic, there was

little overlap between the two stains.


Discussion

This project began with the observation that an antibody (EAAT1-R) against a

known plasma membrane transporter (EAAT1) labeled the nucleus of human fibroblasts.

The EAAT3 isoform, on the other hand, was detected on the plasma membrane and in

vesicles throughout the cytoplasm. The clustering observed using the anti-EAAT3

antibody is consistent with the report by Davis et al. showing punctate fluorescence

throughout the cytoplasm with clustering at the cell surface (Davis et al., 1998).

Although the EAAT1 nuclear localization was unexpected, it was not unprecedented. As

discussed in the Introduction to this chapter, Matthews et al. detected nuclear staining








with the EAAT1 -R antibody in the maternal decidua and placental trophoblast cells of rat

placenta using immunocytochemistry (Matthews et al.. 1998). In other cells of the

placenta, no nuclear localization was detected, but rather, EAAT1 was distributed

throughout the cytoplasm and on the plasma membrane. Other plasma membrane

proteins, such as P-glycoprotein and various growth factor receptors, have also been

detected in the nucleus and nuclear membrane by immunohistochemistry (Stachowiak et

al.. 1996; Baldini et al., 1995).

There are various explanations for the data presented above. First, the EAAT1

glutamate transporter may reside in the nucleus and provide a function that is different

from that of the plasma membrane nutrient transporter. If EAAT1 is serving as an amino

acid transporter in the nucleus, it is unlikely to be mechanistically similar to that on the

cell membrane due to the lack of a Na' gradient across the nuclear membrane. However.

the glutamate transporters also carry other ions such as, K' and H-, and the isoforms.

EAAT4 and EAAT5, have intrinsic chloride channel properties. Perhaps, ion transfer is

the primary function within the nuclear membrane. On the other hand, the lack of nuclear

localization observed with the EAAT1-GFP(N3) and GFP(C3)-EAAT1 fusion proteins.

as well as with both the EAAT1-S and EAAT1-D antibodies, presents evidence against

nuclear localization of the native EAAT1 protein. Although no nuclear localization

signaling sequence has been identified in EAAT1, there are hundreds of nuclear

localization sequence variations which are not highly conserved (Boulikas, 1996).

Another alternative is that the protein recognized by the EAAT I -R antibody

might prove to be a nuclear isoform of EAAT1 with a conserved sequence at the N-

terminus. The fact that the EAAT1-S antibody, generated against a different peptide








sequence, did not stain the nucleus would be consistent with the hypothesis that the

EAAT1-R antibody recognized a different, yet related, protein to EAAT1. Also. if a

truncated form of EAAT1 resides in the nucleus, then it is possible that it would be

recognized by the EAAT1-R but not the EAAT1-S antibody. This later explanation.

however, is not supported by the data from the immunoblot analysis in which the

EAAT I -R antibody recognized a nuclear protein corresponding to the molecular weight

of the intact EAAT1 protein.

The third explanation is that the EAAT1-R antibody is detecting a protein that is

completely different from the EAAT1 glutamate transporter, but shares a few amino acids

in common. The peptide inhibition indicates that the EAAT1-R antibody interaction is

specific, therefore, the unrelated nuclear protein would most likely have at least one

homologous epitope to EAAT1. BLAST database searches did not identify any proteins

other than EAAT/ASCT family members, with significant homology to EAAT1, but such

searches do not always reveal small stretches of common sequence. Also, it would be

quite coincidental that the unknown protein has the same molecular mass as EAATI on

an SDS-PAGE gel. The fact that the EAAT1-R antibody showed very little co-

localization with the EAAT1 -GFP(N3) fusion protein is puzzling. It is surprising that the

EAAT1 -R antibody would recognize an unrelated or homologous protein, and not detect

the expressed EAAT1-GFP(N3). However, the fact that EAAT1-S recognized the

EAAT1-GFP(N3), suggests that the EAAT1 -S antibody is specific for EAAT1, and does

not recognize those transporters or related proteins localized to the nucleus.

Although preliminary immunoblot analysis supports the immunofluorescent

nuclear localization of EAAT1, more extensive studies need to be performed. Various





51


membrane markers need to be used to determine whether or not the EAAT1

immunoreactivity is a result of the contamination of nuclear preparations with other

membranes. Future experiments could also include the isolation of the EAAT1

immunoreactive band from an SDS-PAGE gel for protein sequencing, as well as

immunofluorescence on isolated nuclear fractions.

























Figure 3-1. Extracellular staining of human fibroblasts with EAAT3 antibody. Human
fibroblasts were fixed with 4% PFA and subjected to immunohistochemistry using the
procedures described in the Methods Chapter. Panel A demonstrates the clustering of
EAAT3 transporters on the surface of human fibroblasts using a polyclonal anti-EAAT3
antibody detected with a goat anti-rabbit IgG conjugated to FITC. Panel B shows cells
that were incubated with anti-EAAT3 antibody following preadsorption for 12 h at 4C
with 50 jtg/ml of the corresponding peptide antigen. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
An outline was drawn around one cell to distinguish the periphery. The data shown
represent analysis of 0.2 mrn sections through the cells.






53
























Figure 3-2. Intracellular staining of human fibroblasts with EAAT3 antibody and co-
localization with organelle-specific antibodies. Using the methods described in the
Methods Chapter, human fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with antibodies specific for EAAT3 (A), EAAT3 and KDEL (B),
or EAAT3 and transferring receptor (TfR) (C). Co-localization of the proteins was
assayed by using simultaneously a rabbit polyclonal antibody against EAAT3 detected by
FITC-labeled goat anti-rabbit IgG and mouse monoclonal antibodies against KDEL and
TfR detected by Texas Red labeled goat anti-mouse IgG. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. The data shown represent analysis of 0.2 p.m sections through the cells.




























Figure 3-3. Nuclear staining of human fibroblasts with EAAT1-R antibody and co-
localization with nucleus-specific antibodies. Using the methods described in the
Methods Chapter, human fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with antibodies specific for EAAT1-R (A), EAAT1-R that had
been preadsorption for 12 h at 4C with 50 utg/ml of the corresponding peptide antigen
(B), EAAT1-R and 414 (C), or EAAT1-R and D77 (D). Co-localization of the proteins
was assayed by using simultaneously a rabbit polyclonal antibody against EAAT1-R
detected by FITC-labeled goat anti-rabbit IgG, and mouse monoclonal antibodies against
414 and D77 detected by Texas Red labeled goat anti-mouse IgG. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. An outline was drawn around one cell in each panel to distinguish the
periphery. The data shown represent analysis of 0.2 jtm sections through the cells.



























Figure 3-4. Intracellular staining of human fibroblasts with EAAT1-S and EAAT1-C
antibodies. Using the methods described in the Methods Chapter, human fibroblasts were
fixed with -20C MeOH and subjected to immunohistochemistry with antibodies specific
for EAAT I -S (A) and EAAT1-C (B). The EAAT1-S antibody was detected by an FITC-
labeled goat anti-rabbit IgG, and the EAAT1-C antibody was detected by a Cy3-labeled
goat anti-guinea pig IgG. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 ptm sections through the cells.




























Figure 3-5. Intracellular staining of Hela cells with EAAT1-R and EAAT1-S antibodies.
Using the methods described in the Methods Chapter, Hela cells were fixed with -20C
MeOH and subjected to immunohistochemistry with antibodies specific for EAAT1-R
(A) and EAAT1-S (B). The EAAT1-R and EAAT1-S antibodies were detected by an
FITC-labeled goat anti-rabbit IgG secondary antibody. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
Panel A shows the nuclear staining of multiple cells, whereas panel B is one cell
(outlined) with intracellular vesicle staining. The data shown represent analysis of 0.2
utm sections through the cells.






61









Im
&. 100,OOOg pellet
CONFLUENCE
50 75 100%
203


118

86

w




52


Figure 3-6. Immunoblot analysis of EAAT1 in the nuclear and intracellular membrane
fractions from human fibroblasts. A 30 ug aliquot of the 300 x g nuclear fraction (lane
1) and the 100,000 x g total intracellular membrane fraction (lanes 2-4) was subjected to
SDS-PAGE as described in the Methods Section of Chapter 3. Immunoblot analysis was
performed with a 1:1,000 dilution of EAAT 1 -R antibody and was detected with a
1:10,000 dilution of goat anti-rabbit IgG conjugated to horseradish peroxidase (HRP).
The 50, 75, and 100% labels indicates the degree of cell confluence at the time of cell
lysis and subfractionation. The blot shown is representative of three independent
experiments.
























Figure 3-7. Expression of GFP and GFP-EAAT1 fusion proteins in human fibroblasts.
Human fibroblasts were transfected for 3 h with GFP(N3) only (A), or with the EAAT1-
GFP(N3) (B), and GFP(C3)-EAAT1 (C) fusion proteins according to the lipofectamine
protocol described in the Methods Chapter. Following 24-48 h of expression, cells were
fixed with -20C MeOH and visualized by deconvolution microscopy. Images were
processed from three independent transfections and the fluorescence patterns of expressed
proteins were determined to be reproducible. The data shown represent analysis of 0.2
itm sections through the cells.





64
























Figure 3-8. EAAT1 immunofluorescent staining of human fibroblasts transfected with
EAAT1-GFP(N3). Human fibroblasts were transfected for 3 h with the EAAT1-
GFP(N3) fusion protein according to the lipofectamine protocol described in the Methods
Chapter. Following 24-48 h of expression, cells were fixed with -20C MeOH and
stained with antibodies against EAAT1-R (A) and EAAT1-S (B). Both EAAT1
antibodies were detected with a goat anti-rabbit IgG conjugated to Texas Red and
visualized by deconvolution microscopy. Images were processed from three independent
experiments and the staining was determined to be reproducible. The data shown
represent analysis of 0.2 uim sections through the cells.






66
























Figure 3-9. EAAT1 immunofluorescent staining of PAEC transfected with EAAT1-
GFP(N3). PAEC were transfected for 3 h with the EAAT1-GFP(N3) fusion protein
according to the lipofectamine protocol described in the Methods Chapter. Following 24-
48 h of expression, cells were fixed with -20C MeOH and stained with antibodies
against EAAT1-R (A) and EAAT1-S (B). Both EAAT1 antibodies were detected with a
goat anti-rabbit IgG conjugated to Texas Red and visualized by deconvolution
microscopy. Images were processed from three independent experiments and the staining
was determined to be reproducible. The data shown represent analysis of 0.2 uim sections
through the cells.





68














CHAPTER 4
LYSINURIC PROTEIN INTOLERANCE

Introduction


Lysinuric Protein Intolerance (LPI) is an autosomal recessive disease that is

characterized by a defect in dibasic amino acid transport, as well as, an impaired urea

cycle (reviewed by Simell, 1989). Although cases have been reported worldwide, the

highest prevalence is in Finland, where LPI afflicts 1 in 60,000 to 80,000 people.

Patients have an extremely low tolerance for dietary protein, and symptoms of

hyperammonemia are revealed shortly after weaning infants from the high-fat, low-

protein breast milk. Nausea, vomiting, and diarrhea following meals are early indications

of the disorder. Throughout infancy and childhood, patients show signs of growth

retardation and fail to thrive as a result of protein deficiency. They have enlarged livers

and spleens, muscle hypotonia and hypertrophy, osteoporosis, and approximately 20% of

the patients show varying degrees of mental retardation. The only available treatment is

to limit the consumption of protein. Normalization of hepatic nitrogen utilization and

urea synthesis requires supplementing meals with 3 to 8 grams of citrulline daily.

LPI was originally defined by elevated urinary levels and poor intestinal

absorption of all cationic and several neutral amino acids (Simell, 1989). Normal daily

urine contains a mean of 4.13 mmol lysine/1.73 m2 body surface area, whereas the urine

from LPI patients contains a mean of 25.7 mmol lysine/1.73 m2 body surface area








(Simell, 1989). The concentrations of cationic amino acids in plasma are low, whereas

glutamine, alanine, serine, proline, citrulline, and glycine are increased. Although

cationic amino acid transport is probably defective in most cell types, it has been

documented to be so in kidney tubules, intestine, and cultured fibroblasts (Simell. 1989).

LPI-derived fibroblasts accumulate elevated steady state levels of cationic amino acids

and exhibit a reduced rate of trans-membrane exchange for these same substrates (Smith

et al., 1987). The biochemical basis for this decreased release of cellular amino acids has

not been established. Although several amino acidurias are thought to result from

defective Na'-dependent transport across the brush border domain of either intestinal or

renal epithelial cells, previous studies have suggested that the primary transport defect in

LPI is a reduced efflux across the basolateral surface (Simell, 1989; Rajantie et al., 1981).

Evidence for this interpretation comes from an observation that plasma amino acid

concentrations remain low following the oral administration of both cationic amino acids

and lysine dipeptides. Dipeptides are transported normally across the brush border

membrane by a mechanism distinct from that of free amino acids. The dipeptides are

then hydrolyzed to free amino acids by intracellular enzymes. However, in LPI

enterocytes, these free amino acids are unable to pass through the basolateral membrane

into the plasma, and instead, are transported back out via a brush border membrane

transporter. Direct measurements of lysine transport in intestinal biopsy specimens have

confirmed that the transport defect is located on the basolateral membrane (Simell, 1989).

Transport appears to be normal at the luminal surface of renal epithelial cells from LPI

patients (Simell, 1989).











Lysine Transport in Whole Cells and Plasma Membrane Vesicles of LPI

System y- is a Na--independent activity that catalyzes facilitated transport and

exchange reactions (White, 1985), but also may permit accumulation of cationic amino

acids against a concentration gradient driven by the trans-membrane potential (Bussolati

et al., 1987). Under normal physiological conditions. System y- mediates the uptake of

cationic amino acids such as arginine, lysine, and ornithine. CAT1 was the first cDNA to

be identified that exhibited System y- activity (Albritton et al., 1989). This transporter.

which also serves as the murine ecotropic leukemia virus receptor, mediates the Na-

independent, high-affinity transport of cationic amino acids in most mammalian cells

(Kim et al., 1991; Wang et al., 1991). The second member of the "CAT" family to be

isolated, CAT2, shares 61% identity with CAT1 at the amino acid level, and mediates

Na'-independent, high-affinity transport of cationic amino acids in activated T

lymphocytes (MacLeod et al., 1990). The third clone to be identified, CAT2a, is the

result of alternative splicing of the CAT2 gene. It differs from CAT2 by only 41 amino

acids, yet it exhibits a 10-fold lower affinity for arginine and is liver-specific (Closs et al..

1993). The last member of the "CAT" family, CAT3, shares the greatest homology with

CAT1 and mediates the Na'-independent high-affinity transport of cationic amino acids

in the brain (Hosokawa et al., 1997; Ito and Groudine, 1997).

In cultured fibroblasts, there is little difference between the initial rates (measured

after 15 sec) of lysine uptake in normal and LPI cells, however, after 20 min, the LPI

cells accumulate two-fold more 3H-lysine than control cells (Handlogten and Kilberg,








unpublished results). The lack of a large difference in initial uptake rates suggests that

influx at the plasma membrane is unaffected by the disease. To eliminate the possible

confounding effects oftrans-stimulation when measuring transport in whole cells, a crude

mixture of cellular membrane vesicles was isolated from cultured fibroblasts and used to

assay lysine uptake. Qualitatively, the data obtained from vesicle transport experiments

were similar to the results of whole cell transport. Once again, initial measurements

showed little difference in uptake by normal and LPI-derived vesicles, whereas LPI-

derived vesicles eventually accumulated four to five times more 3H-lysine than the

control vesicles (Handlogten and Kilberg, unpublished results). These results suggested a

decreased efflux of amino acid from either the plasma membrane or intracellular vesicles

of the LPI cells.

The observation that the LPI-derived vesicles also accumulate select neutral

amino acids (Handlogten and Kilberg, unpublished data) is consistent with the detection

of increased neutral amino acids in the urine of LPI patients. This prompted our

laboratory to investigate amino acid transport by other known transport systems in LPI

fibroblasts and plasma membrane vesicles. System b" is a Na'-independent system that

mediates the bidirectional transport of both cationic and neutral amino acids (Van

Winkle, 1988; Van Winkle et al., 1988). However, this system is probably not

responsible for the elevated Na--independent uptake of neutral amino acids observed

during in vitro studies, because the transport of leucine was poorly inhibited by lysine in

the plasma membrane vesicles derived from LPI fibroblasts (Handlogten and Kilberg,

unpublished results). The NBAT protein, which exhibits the properties of System b",,








has been implicated as the defective cystine transporter in the inherited disease, cystinuria

(Calonge et al.. 1994).

System B'-. first described by Van Winkle and colleagues (Van Winkle et al..

1985), is responsible for cationic and neutral amino acid transport in various cell types.

including human fibroblasts. However, this is a Na'-dependent system, and therefore.

does not appear to be responsible for the increased LPI amino acid accumulation. System

yL mediates the Na'-independent high-affinity transport of cationic amino acids, as well

as the Na--dependent high-affinity transport of L-leucine (Deves et al., 1992). The

observation by our laboratory that 5 mM 2-amino-[2,2,1]-bicycloheptane-2-carboxylate

(BCH) inhibited leucine uptake in the normal and LPI vesicles is consistent with the

hypothesis that System y'L may be responsible for the unusual accumulation of cationic

and neutral amino acids in the LPI vesicles. L-leucine transport was inhibited by leucine,

lysine, arginine. and BCH in normal fibroblasts vesicles. However, lysine did not block

the L-leucine uptake in LPI vesicles. Collectively, the vesicle transport data is not

entirely consistent with the properties of any of the known transport systems. Therefore,

it is possible that an uncharacterized transport system is responsible for the activity

observed in the LPI vesicles. Alternatively, a known transport system that is altered in its

substrate specificity in the LPI cells may account for the elevated amino acid

accumulation.

Immunofluorescence Studies

When transport studies in LPI cells began, CAT1 was the only family member

that had been reported to mediate System y' activity. Therefore, based on clinical







observations in LPI patients, it was originally hypothesized that LPI may arise from a

defect in the cationic amino acid transporter, CAT1, previously termed System y' (Smith

et al., 1987). However, two independent laboratories sequenced the human CAT 1

mRNA expressed in LPI patients and found no mutations within the sequence (personal

communication, Dr. Olli Simell, University of Turku). As a result, our laboratory

hypothesized that a trafficking defect involving the CAT 1 transporter, or a protein

involved in membrane protein trafficking, may be responsible for the altered transport of

cationic amino acids in LPI cells. Preliminary immunofluorescence studies in our

laboratory revealed a population of intracellular vesicles, unique to LPI cells, which

appeared to contain the CAT1 transporter (Woodard and Kilberg, unpublished data). The

elevated steady-state accumulation of lysine, described above, supports the hypothesis

that amino acids may be sequestered in the abnormal intracellular vesicles of LPI cells. If

the defect involves a deficiency in efflux across the basolateral membrane domain in

epithelial cells, as proposed, perhaps the reason is that the lysine becomes trapped within

the intracellular vesicles instead of rapidly equilibrating across the plasma membrane.

This hypothesis is consistent with the finding by Smith et al. that trans-stimulation of

cationic amino acid efflux is also decreased in LPI cells (Smith et al., 1987).

Dr. Olli Simell, at the University of Turku (Turku, Finland), has been following

Finnish patients with LPI for over 25 years. His laboratory has thoroughly documented

the clinical aspects of the disease and is currently working to localize and characterize the

defective gene. Using linkage analysis and a candidate gene approach, Simell and

coworkers have excluded the possibility that CAT1 or CAT2 (Lauteala et al., 1997) is the

defective gene. Despite these observations, it is clear that the LPI fibroblasts exhibit








abnormal cationic and neutral amino acid transport and accumulate abnormal intracellular

vesicles and vacuoles. Therefore, this project has focused on the evaluation of membrane

protein trafficking in the cells of LPI patients.


Trafficking Defects in Plasma Membrane Transport Proteins

There is precedence for the aberrant trafficking of a specific class of transporter

proteins in both Saccharomyces cerevisiae and Drosophila. It was shown that S.

cerevisiae requires an endoplasmic reticulum (ER) integral membrane protein, SHR3, for

the effective processing and trafficking of amino acid permeases, specifically (Ljungdahl

et al., 1992). Mutations in SHR3 cause retention of 13 out of 13 amino acid permeases

tested within the ER, and therefore, block amino acid uptake by interfering with the

plasma membrane localization. The defective SHR3 does not affect the targeting of any

other plasma membrane, secretary, or vacuolar protein (Ljungdahl et. al., 1992). In

Drosophila, mutations in a gene (ninaA) encoding an ER cis-trans isomerase cause the

abnormal intracellular accumulation of two homologous opsin proteins in photoreceptor

cells. The accumulation is presumed to occur as a result of improper protein folding early

in the biosynthetic pathway (Colley et. al., 1991).

In humans, a single amino acid deletion in a highly regulated chloride channel is

the basis for the fatal genetic disorder cystic fibrosis (CF). It was shown that the

mutation that was present in 70% of the defective CF genes (Kerem et al., 1989) resulted

in the abnormal retention of the cystic fibrosis transmembrane conductance regulator

protein (CFTR) in the endoplasmic reticulum, although the Cl- conductance activity

measured in ER vesicles or reconstituted proteoliposomes was unimpaired (Cheng et al.,








1990: Drumm et al., 1991). Whereas both the wild-type and mutant CFTR proteins are

associated with calnexin in the ER. only the wild-type protein exits the ER and is

correctly trafficked to the plasma membrane (Pind et al.. 1994).

Although no mutations have been found in the CAT1 transporter gene in LPI

patients, or any other protein at this point, the LPI defect results in pleiotropic effects on

amino acid transport (Simell, 1989). Broad scope changes in transport could result from

a block of transporter processing at any stage along the various pathways of membrane

protein synthesis or endocytic recycling. This chapter describes the cellular localization

of the CAT1 arginine transporter and of various organelle-specific proteins along the

biosynthetic, endocytic, and degradative pathways in normal and LPI fibroblasts.

Results

Morphology of normal and LPI fibroblasts by light and electron microscopy.

There are basic differences between the normal and LPI cells that can be discerned at the

light microscope level (Figure 4-1). The LPI cells frequently appear to be larger with

long processes and extensions at the cell periphery. Also, normal fibroblasts grow

significantly faster than their LPI counterparts under the same culture conditions until

they reach confluency. The LPI cells tend to grow in clusters and never reach full

confluence. The most striking difference between the normal and LPI cells is a

population of large vesicles or vacuoles observed throughout the cytoplasm, often

concentrated around the nucleus of the LPI cells. When electron microscopy was used to

increase the resolution of these large intracellular vacuoles, at a magnification of 25K

they appear to contain a fibrous material of unknown origin (Figure 4-2). The unusual








LPI-derived vacuoles have been detected by electron microscopy in the cell lines from

two different LPI patients (Woodard and Kilberg, unpublished data: McDonald and

Kilberg, unpublished data). The only previous microscopic observation reported was in a

paper by Simell and coworkers who mentioned apparent alterations in the architecture of

hepatocytes (Simell et al., 1975). They reported an accumulation of vesicles containing a

"fibrillogranular material," although it is not known whether these represent the same

vesicles we have identified.

CAT1 antibody production against human. The preliminary immunofluorescence

research was performed using a CAT1 polyclonal antibody that was generated against a

25 amino acid sequence (SIKNWQLTEKNFSCNNNDTNVKYGE) from the third

extracellular loop of the murine CAT1 sequence (Woodard et al., 1994). The murine

CAT1 antibody was shown to stain specifically a wide variety of cell types from several

species and was inhibited by pre-incubation with the corresponding peptide. However,

this anti-murine CAT1 antibody did not immunoblot well, and it is possible that it is not

optimal for detection of denatured protein in the human cell lines being tested. For this

reason, a polyclonal antibody was generated against a sequence of 20 amino acids

(CEEASLDADQARTPDGNLDQ) at an intracellular site of the human CAT1 protein

(Cocalico Biologicals, Inc., Reamstown, PA). Enzyme-linked immunosorbent assays

(ELISA) were performed on the individual serum samples to determine antibody tigers.

Briefly, 96-well plates were coated with the human CAT1 peptide (above), and incubated

with a 1:50 to a 1:32,000 dilution of the immune or pre-immune serum collected from the

inoculated rabbit. Following the removal of the primary antibody, an alkaline-

phosphatase conjugated secondary antibody was added to the plates and detected with a








phosphatase substrate at a wavelength of 405 runm. By the sixth bleed, one of the human

CAT1 antibodies showed a 2.5-fold increase in absorbance over the pre-immune at a

dilution of 1:50. Although a series of immunohistochemistry experiments were

conducted using this anti-human CAT1 antibody, the background staining was high and

the corresponding peptide failed to completely inhibit staining. Therefore, all of the data

presented in this chapter was generated using the polyclonal murine CAT1 antibody. The

conditions for the CAT1 antibody as well as the other antibodies used in this chapter are

summarized in Table 4-1.


Name

CAT1

EAAT1

EAAT3

GLUT1

414

D77

P3-integrin

caveolin

tubulin

KDEL

BiP

PDI

sialyltransferase


Table 4-1
Antibodies for Immunofluorescence Studies
Host Source

rabbit Dr. Michael Kilberg,
Univ. of Florida
rabbit Dr. Jeffrey Rothstein.
John Hopkins
rabbit Dr. Michael Kilberg,
Univ. of Florida
rabbit Dr. Susan Frost,
Univ. of Florida
mouse Dr. John Aris,
Univ. of Florida
mouse Dr. John Aris,
Univ. of Florida
mouse Dr. Martin Hemler,
Dana-Farber Cancer Inst.
rabbit Transduction Laboratories
Lexington, KY
mouse Sigma,
St. Louis, MO
mouse Dr. David Vaux,
EMBL
rabbit Dr. Susan Frost,
Univ. of Florida
rabbit Dr. Tom Wileman,
Dana-Farber Cancer Inst.
rabbit Dr. William Dunn,
Univ. of Florida


Dilution

1:25

1:50

1:200

1:50

1:10

1:50

1:100

1:200

1:100

1:100

1:50

1:25

1:50








Name Host Source Dilution

M6P receptor rabbit Dr. Peter Nissley, 1:50
NIH
Rab 5 rabbit Santa Cruz Biotechnology 1:100
Santa Cruz, CA
transferring receptor mouse Zymed, 1:5
San Francisco. CA
lpg 120 Rabbit Dr. William Dunn, 1:100
Univ. of Florida
cathepsin D Rabbit Biodesign International, 1:300
Kennebunk, ME


Distribution of endogenous CAT1 in normal and LPI fibroblasts. As mentioned

above, the LPI disease was originally characterized by elevated concentrations of arginine

and lysine in the urine. Later, it was determined that the transport defect is expressed in

the kidney tubules, intestines, cultured fibroblasts, and probably hepatocytes (Simell,

1989). It has been shown that CAT2 and CAT2a are not expressed in human fibroblasts,

so our laboratory began to investigate the localization of the CAT1 transporter in normal

and diseased cells. Normal and LPI fibroblasts were fixed with 4% PFA and labeled with

a 1:25 dilution of the CAT1 transporter antibody (according to the protocol in the

Methods chapter). Antibody staining was detected with a 1:200 dilution of goat anti-

rabbit IgG linked to FITC. Both normal and LPI fibroblasts demonstrated an

extracellular periodic labeling that resembled intensely stained patches on the plasma

membrane (data not shown). Staining of the fluorescent patches was completely blocked

when the murine CAT1 antibody was pre-incubated with 50 ig/ml of corresponding

peptide for 12 hours before labeling. This same pattern was observed in several different

LPI fibroblast cell lines, as well as, in porcine pulmonary artery endothelial cells (see

Chapter 5). Incubation of the cells with the microtubule inhibitor, nocodazole, caused the








transporter staining to disperse over the entire cell surface, and removal of the inhibitor

caused the clusters of transporter to reform within 3 h (data discussed in Chapter 5).

These data confirmed previous reports by Woodard et al. that the CAT1 antibody forms

clusters on the plasma membrane of normal fibroblasts, and that this arrangement is

dependent on intact microtubules (Woodard et al., 1994). Identification of these clusters

as caveolae will be discussed in Chapter 5.

In a separate series of experiments, the intracellular distribution of the CAT1

transporter was detected by immunofluorescence and deconvolution microscopy in

MeOH-fixed fibroblasts. Although an intracellular vesicle population was labeled with a

1:25 dilution of the CAT1 antibody in both normal (Figure 4-3 A) and LPI fibroblasts

(Figure 4-3 B), there was not a significant difference in the number or physical

appearance of the vesicles. Incubating cells with secondary antibody only, or incubating

the primary antibody with 50 jig/ml of corresponding peptide, resulted in only a faint,

diffuse background immunofluorescence (data not shown). These results are different

than earlier reports from our laboratory that documented an obvious difference between

the normal and LPI cells when observed using an epifluorescence microscope. It was

observed previously that CAT1 antibody labeling of normal fibroblasts resulted in diffuse

cytoplasmic staining with slightly increased intensity in the region of the Golgi complex

(a perinuclear pattern). On the other hand, the CAT1 antibody staining in the LPI cells

was specific for a large number of intracellular vesicles, which appeared to be randomly

distributed throughout the cytoplasm. As a result of the apparent discrepancy, staining of

normal and LPI fibroblasts was repeated using a variety of conditions. CAT1 serum.








IgG-purified, and affinity-purified antibodies were used at dilutions ranging from 1:25 to

1:1000. In addition, cells were fixed with either -20C MeOH or 2%-4%

paraformaldehyde, for 10 to 30 minutes, before permeabilizing and incubating with the

CAT1 antibody. Several different normal and LPI patient cell lines were used and

antibody incubations ranged from 1 h to overnight (data not shown). However, none of

these conditions produced the CAT 1-containing LPI vesicles seen previously. Additional

research, perhaps with new antibody preparations, will be required to confirm the CAT1

staining of LPI fibroblasts.

Previous LPI research in the Kilberg laboratory was limited to the study of the

CAT1 transporter, yet it is entirely possible that the proposed trafficking defect involves

other transporter, and/or non-transporter, proteins. Antibodies against several other

nutrient transporters were available and used for intra- and extracellular

immunofluorescent labeling of normal and LPI fibroblasts. Both cell types were stained

with a 1:50 dilution of the Na-independent glucose transporter, GLUT1, a 1:200 dilution

of EAAT3, and a 1:50 dilution of EAATI following both -20C MeOH and 4%

paraformaldehyde fixation. Each of these antibodies was detected using a goat anti-rabbit

IgG conjugated to FITC. GLUT1 labeling was detected in small amounts on the plasma

membrane, but primarily in cytoplasmic and perinuclear vesicles. The EAAT3 antibody

was detected in clusters on the plasma membrane, as well as in vesicles throughout the

cytoplasm. The EAAT1 antibody, discussed in detail in Chapter 3, was localized to the

nuclear membrane. No differences in the staining patterns were detected, and no

abnormal intracellular vesicle populations were identified, when labeling the normal and

LPI cells with antibodies against any of these transporters (data not shown).








Localization of expressed CAT1 in normal and LPI fibroblasts. The experiments

investigating endogenous CAT1 localization relied on the detection of the human

transporter with a anti-murine CAT 1 antibody. An expression system was developed in

order to confirm endogenous experiments, as well as, to avoid any technical problems

associated with the CAT1 antibody. Transfection with several CAT1 constructs was

attempted before deciding to use the green fluorescent protein (GFP) vector from

Clontech (Palo Alto, CA). These constructs included a FLAG-tagged murine CAT1

cDNA, generated in our laboratory, as well as an HA-tagged human CAT1 cDNA and a

human CAT1 -GFP fusion protein, both provided by Dr. Lorraine Albritton at University

of Tennessee Medical Center. Neither the HA-tagged CAT1 nor the FLAG-tagged CAT1

constructs were detected following transfection, and the expression of the CAT1-GFP

from Dr. Lorraine Albritton was extremely low (data not shown). Dr. Lorraine Albritton's

CAT1-GFP fusion protein was constructed with a GFP variant that was not modified for

enhanced expression and fluorescence in human cells. Therefore, a GFP(C3)-CAT1

fusion protein was constructed using Clontech's humanized GFP variant and the CAT1

cDNA sequence from mouse (See Methods Chapter).

The GFP tag (described in the Methods Section) provided an alternative technique

for observing the localization of the CAT1 protein in the normal and LPI fibroblasts. If

the CAT1 I transporters were trafficking incorrectly in LPI cells, but the available antibody

could not detect it, then the GFP(C3)-CAT1 would provide an independent method of

CAT 1 localization. Transfection of normal fibroblasts with the GFP(C3) only, which

lacks a targeting signal and, therefore, diffuses throughout the cytoplasm and nucleus,

was used as a control (Figure 4-4 A). Alternatively, transfection of normal fibroblasts








with the GFP(C3)-CAT1 demonstrated specific targeting to the plasma membrane, as

well as one or more intracellular vesicle populations (Figure 4-4 B). The strong

perinuclear staining is probably GFP(C3)-CAT1 in the Golgi. The expressed GFP(C3)-

CAT1 fusion protein also showed significant co-localization with the CAT1 antibody

(detected with a 1:200 dilution of goat anti-rabbit IgG linked to Texas Red) in normal

fibroblasts (Figure 4-4 C). Although the GFP(C3)-CAT1 made it easier to visualize the

exact location of the CAT1 transporter, there was no difference in the intracellular pools

of GFP-CATI in normal (Figure 4-5 A) and LPI (Figure 4-5 B) fibroblasts. In both cell

lines, the GFP-CAT1 fusion protein was detected on the plasma membrane, throughout

the cytoplasm in small vesicles, and highly concentrated in the perinuclear region

(probably representing Golgi).

Examination of organelle integrity in normal and LPI fibroblasts. As discussed,

no significant differences in the distributions of any of the amino acid transporters were

identified in the normal and LPI fibroblasts. However, given the differences in

morphology observed between normal and LPI fibroblasts and the presence of the LPI-

specific vacuoles, it is clear that the disorder is associated with a basic cellular defect that

can be visualized as a structural deformity in one or more of the organelles of the LPI

cells. Therefore, staining patterns generated by organelle-specific antibodies, in normal

and LPI cells, were compared in order to gain important information regarding the

integrity of the organelles in the LPI fibroblasts. A variety of organelle-specific

antibodies were used as markers for the identification of compartments involved in the

biosynthetic, endocytic, and degradative pathways. The antibodies that were chosen are

prototypical markers for membrane protein trafficking compartments and have been








described in the literature for a variety of cell lines. Initial experiments were performed

in order to confirm the reliability of the antibodies and document the staining patterns in

the normal human fibroblasts. For all of the immunoassays, secondary antibodies used

were either goat anti-rabbit IgG or goat anti-mouse IgG conjugated to either fluorescein

isothiocyanate (FITC, Sigma Chemical Co., St. Louis, MO) or Texas Red (TR, Cappel

Laboratories, Durham, NC).

The nuclear membranes of normal (Figure 4-6 A) and LPI (Figure 4-6 C) cells

were examined using a 1:10 dilution of mouse anti-human 414 antibody, which was

generated against an epitope common to several of the nuclear pore complex proteins

(Davis and Blobel, 1986). A 1:50 dilution of mouse anti-yeast D77 antibody (Aris and

Blobel, 1988) was used to visualize the nucleoli of both normal (Figure 4-6 B) and LPI

(Figure 4-6 D) fibroblasts. The labeling of the nuclear structures was consistent with

images from the literature, and there was no discernible difference between the staining

patterns of the normal and LPI fibroblasts. A 1:100 dilution of mouse anti-human-(3-

integrin antibody was used to label the plasma membrane (Figure 4-7 A and B), and a

1:200 dilution of rabbit anti-caveolin-1 antibody was used to specifically detect the

caveolar domains of the plasma membrane (Figure 4-7 C and D). In both normal and LPI

cells, the caveolin antibody labeled specific regions of the plasma membrane, as

expected, and the P-integrin antibody showed strong staining around the periphery of

both cell types. Although the LPI cells may be defective in cell-to-cell contact, indicated

by the way they grow in clusters rather than spreading out to confluence, there was no

noticeable difference in the P-integrin staining pattern of the normal and LPI cells. The








microtubules appeared to be intact in both the normal (Figure 4-7 E) and LPI cells

(Figure 4-7 F) according to immunofluorescent labeling using a 1:100 dilution of mouse

anti-p3-tubulin antibody.

Antibodies against resident proteins of the endoplasmic reticulum (ER) and Golgi

were used to compare the organelles of the biosynthesis pathway in normal and LPI cells.

A short, four amino acid (Lys-Asp-Glu-Leu) peptide, called KDEL, appears in the

sequences of ER resident proteins (such as BiP) and is responsible for selectively

retrieving the proteins after they leave the ER in transport vesicles (reviewed by Pelham,

1991). A membrane-bound receptor in the cis-Golgi recognizes the KDEL retention

signal, and returns the KDEL-containing proteins to the ER. The ER, in normal and LPI

fibroblasts, was detected using a 1:100 dilution of mouse anti-KDEL antibody (Figure 4-

8 A and B), a 1:50 dilution of rabbit anti-BiP antibody (data not shown), as well as, a

1:25 dilution of rabbit anti-protein disulfide isomerase (PDI) antibody (data not shown).

The KDEL antibody provided the best labeling, but each of the antibodies stained the ER

with no significant difference between the two cell lines. The ER in the LPI cells

appeared to be larger and more spread out than the ER of the normal fibroblasts;

however, this was probably due to the fact that the LPI cells tend to be larger in general.

Compartments of the Golgi were stained using 1:50 dilutions of either rabbit anti-

sialyltransferase antibody (Figure 4-8 C and D), which labels the Golgi and Trans-Golgi

Network (TGN), or rabbit anti-mannose-6-phosphate receptor antibody (Figure 4-8 E and

F), which is specific for TGN and late endosomes. There was no recognizable difference








in the structure of the Golgi, or any component of the biosynthetic pathway that was

investigated in the normal and LPI fibroblasts.

Organelles of the recycling pathway were visualized using a 1:100 dilution of

rabbit anti-human Rab5 antibody (Figure 4-9 A and B), for labeling early endosomes, and

a 1:5 dilution of mouse anti-human transferring receptor antibody (Figure 4-9 C and D),

for detecting vesicles involved in endocytosis and recycling. Neither of these antibodies

detected an abnormality in the recycling pathway of the LPI cells. The number of

vesicles containing the transferring receptor varied significantly between experiments, but

the variation was independent of normal versus LPI. Vesicles and compartments

involved in degradation were identified using a 1:300 dilution of the lysosomal protease,

cathepsin D (Figure 4-10 A and B). Although the antibody demonstrated a punctate

staining pattern in both cell types, the vesicles detected in the LPI fibroblasts (Figure 4-10

B) were larger and in greater abundance than those observed in the normal fibroblasts

(Figure 4-10 A). To further test for a difference in the lysosomal staining, normal and

LPI cells were labeled with a 1:100 dilution of a mouse antibody generated against the

lysosomal membrane protein, lpgl 120 (Figure 4-11 A and B). In the normal fibroblasts

(Figure 4-11 A), a punctate pattern of lysosomal labeling was observed throughout the

cytoplasm. On the other hand, the apparent diameter of the lysosomes detected in the LPI

fibroblasts (Figure 4-11 B) were larger and in greater abundance than in the normal cells.

In addition, the LPI lysosomes were tightly clustered and located in close proximity to the

nucleus.

Treatment of normal and LPI fibroblasts with lysosomotropic agents. From the

labeling, it appeared as though the lpg 120-containing lysosomes corresponded to the








large vacuoless" that were previously observed by phase contrast and electron

microscopy. Acridine orange (AO) is a lysosomotropic weak base that accumulates in

acidic compartments of living cells (Robbins et al., 1963). When excited with blue light,

it emits a red fluorescence that can be visualized using a Texas Red filter. To determine

whether or not the lpg 120-containing compartment was acidic in nature, as would be

expected for lysosomes, normal and LPI fibroblasts were loaded with 5 p.g/ml acridine

orange for 15 min before fixing with either 4% paraformaldehyde or-20C MeOH and

viewing with a Nikon axiophot inverted epifluorescent microscope (Figure 4-12 A and

B). Deconvolution microscopy could not be performed because the acridine orange

diffused out of the acidic compartments too rapidly. For the same reason, staining with

the lpg 120 antibody after loading with acridine orange was unsuccessful. However, it did

appear as though the AO-containing compartment in the LPI cells was larger and more

abundant (Figure 4-12 B), even though it was not confirmed that these structures

represented the lpg 120-containing lysosomes. The same AO staining patterns were

obtained when live normal or LPI cells were loaded with AO and immediately viewed

with the epifluorescence microscope. When exposed to an extremely high dose of AO

(ranging from 50-500 p.g/ml), the normal fibroblasts died immediately, as judged by the

way they curled and lifted off the tray, whereas the LPI cells were able to tolerate the

drug (data not shown).

The lysosomes of normal and LPI fibroblasts were also compared following

treatment with chloroquine, a lysosomotropic drug that neutralizes lysosomes and other

acidic compartments. Both cell lines were incubated with 50 pM of chloroquine in MEM








+ 10% FBS for 1 h at 37C before fixing with -20C MeOH and staining with anti-lpgl20

antibody (Figure 4-13 A-D). The normal fibroblasts, without chloroquine treatment,

contained small lysosomes distributed throughout the cytosol, as visualized by labeling

with the lpg 120 antibody (Figure 4-13 A). However, following chloroquine treatment,

the lpg 120-containing compartment in the normal fibroblasts (Figure 4-13 B) increased in

size and number resembling the lysosomal-like vacuoles of the untreated LPI cells

(Figure 4-13 C). When the LPI fibroblasts were treated with chloroquine, the lpgl20-

containing vacuoles grew to an enormous size and clustered tightly in the perinuclear

region of the cells (Figure 4-13 D). Incubating both cell types with a higher

concentration of chloroquine, or for an extended period of time resulted in similar, yet

even more pronounced effects. The vacuoles in the normal fibroblasts became even

larger and more abundant, whereas the intracellular membranes of the LPI cells appeared

to dissolve until only a few giant vacuoles filled the entire intracellular space (data not

shown).

Antibody labeling of normal and LPI fibroblasts expressing GFP-CAT1. The

experiments discussed in this section are an extension of those presented above. As

mentioned before, no differences were detected in the abundance or distribution of

endogenous or exogenous CAT1 in normal and LPI fibroblasts. Of all of the organelle-

specific antibodies tested, only those associated with the lysosomes showed any

abnormalities in the LPI cells. The following co-localization experiments examined

whether or not the CAT1 transporter, in LPI cells, was trapped in a particular

compartment or vesicle pool, such as the lysosomes. Normal and LPI fibroblasts were

transfected with the GFP-CAT1 fusion protein according to the lipofectamine protocol








described in the Methods Chapter. After 24 h, the cells were fixed with -20C MeOH,

and stained with organelle-specific antibodies (according to normal transfection and

immunofluorescence procedures described in the Methods Chapter). The GFP-CAT1

transfection was used instead of the CAT1 antibody because many of the organelle-

specific antibodies were generated in the same species as the CAT1 antibody, and

therefore, could not be used for double-labeling experiments. The Rab5 antibody was

used to co-localize GFP-CAT1 with early endosomes, the mannose-6-phosphate receptor

antibody was used to detect GFP-CAT1 in Golgi and late endosomes/TGN, and the

lpg 120 antibody was used to observe GFP-CAT1 associated with the late endosomes and

lysosomes. The results obtained were the same for experiments conducted in both normal

and LPI fibroblasts. The GFP-CAT1 co-localized to a moderate degree with Rab5 and

mannose-6-phosphate receptor antibodies (data not shown); however, very little overlap

was observed with the lpgl 120 antibody (Figure 4-14 A and B). This was a reasonable

result given the probable participation of the transporter in the biosynthetic and recycling

pathways. Even though the lysosomes, or a lysosome-like vesicle population, appear to

be abnormal in the LPI cells, there was no accumulation of the GFP-CAT1 in these

compartments.


Discussion

Previous experiments demonstrated that LPI fibroblasts exhibited an elevated

accumulation of cationic amino acids over time, as well as the existence of a unique

vesicle population. As a result, it was hypothesized that the CAT1 transporter may be

trapped in an intracellular compartment and, therefore, unable to mediate efflux of








cationic amino acids across the basolateral plasma membranes of affected cells. This

transporter defect would provide an explanation for the increased intracellular

accumulation of cationic amino acids during LPI whole cell and vesicle transport assays.

Experiments were performed to test for the presence of CAT 1, or other transporters, in

abnormal LPI vesicles using a higher-resolution microscope than had previously been

available. In addition, organelle-specific antibodies were used to compare the cellular

compartments involved in biosynthesis, endocytosis, and degradation in normal and LPI

cells. Immunofluorescence studies did not detect an association of CATI transporters

with the abnormal LPI-associated vesicle population previously reported. In addition, the

extracellular and intracellular distributions of the CAT 1, EAAT1, EAAT3, and GLUT1

transporters, as well as the integrity of specific organelles involved in the trafficking

pathways of membrane proteins appeared normal in the LPI fibroblasts. Therefore, the

most important difference between the normal and LPI fibroblasts, detected by

immunofluorescence in this study, was an abnormal population of enlarged lysosomal-

like vacuoles (described above).

It is confusing as to why earlier immunofluorescence data documenting the

CATI-containing LPI vesicles could not be reproduced. However, several experimental

conditions were changed that may explain the apparent discrepancy. First, although the

primary cell line used in this study came from the same patient as before, the cells were

isolated and cultured at different times. Although it is unlikely, the presence of CAT1 in

the LPI vesicles could have involved a defect specific to the cell line used in the first set

of experiments. Second, the same sample of CAT 1 antibody was not available for this

study. The CAT 1 peptide and antibody were older during the experiments described




Full Text
BIOGRAPHICAL SKETCH
Kelly Kristin McDonald was bom to Maurice and Patricia McDonald on August
1, 1971, in Nashville, Tennessee. After graduating from John Overton High School, in
Nashville, Tennessee, she attended Western Kentucky University where she studied
theatre and dance as a Performing Arts major for three years. In 1992, Kelly transferred
to the University of Florida and earned a B. S. in Biochemistry. She began graduate
studies in the Department of Biochemistry and Molecular Biology at the University of
Florida in January 1994. Following graduation, Kelly will begin work on a Master's in
Scientific Journalism and Mass Communications at the University of Florida. In
addition, she will continue to conduct research with Dr. Edward Block at the University
of Florida College of Medicine.
185


121


TABLE OF CONTENTS
ACKNOWLEDGMENTS ii
LIST OF TABLES v
LIST OF FIGURES vi
ABSTRACT ix
CHAPTER 1 INTRODUCTION 1
Overview of Mammalian Amino Acid Transport 1
Trafficking of Membrane Proteins 11
Cell Biological Techniques for Studying Protein Trafficking 15
CHAPTER 2 MATERIALS AND METHODS 17
Materials 17
Methods 18
CHAPTER 3 DISTRIBUTION OF THE GLUTAMATE TRANSPORTERS 28
Introduction 28
Methods 36
Results 39
Discussion 48
CHAPTER 4 LYSINURIC PROTEIN INTOLERANCE 69
Introduction 69
Results 76
Discussion 89
CHAPTER 5 CAVEOLAR COMPLEX BETWEEN THE CATIONIC AMINO
ACID TRANSPORTER 1 AND ENDOTHELIAL NITRIC
OXIDE SYNTHASE 124
Introduction 124
Methods 129
Results 132
iii


A


13
recently, have been implicated in targeting certain proteins to plasma membrane caveolae
(Song et al., 1996). Following translation, proteins are packaged into vesicles and
transported to the Golgi apparatus for the completion of glycosylation and folding events.
The developing proteins are shuttled from the cis, to the medial, to the trans-Golgi
compartment, ultimately arriving at the trans-Golgi network (TGN) where they are sorted
according to their final destinations (Alberts et al., 1994).
Membrane proteins also participate in the endocytic/exocytic pathway. Several
distinct forms of endocytosis have been described (reviewed by Watts and Marsh, 1992;
Alberts et al., 1994). During pinocytosis, small invaginated plasma membrane vesicles of
less than, or equal to 150 nm in diameter, constituatively carry fluids and solutes into the
cell. Phagocytosis, on the other hand, results in the regulated ingestion of large particles
via plasma membrane derived vesicles of greater than 250 nm in diameter, and is
generally the responsibility of specialized cells. It is assumed that there is an
internalization of many plasma membrane proteins during these two processes. Most
animal cells take up specific macromolecules by a process called receptor-mediated
endocytosis. During this event, receptor-ligand complexes participating in this cycle are
internalized by clatherin-coated pits on the plasma membrane and delivered to early
endosomes. The acidic environment of the endosme results in the dissociation of the
ligands, which advance via late endosomes to ultimate lysosomal degradation. Some
receptors are recycled directly back to the plasma membrane from the endosomal
compartment, whereas others recycle by way of an intermediate step in the TGN. The
transferrin receptor (TfR.) interacts with iron-bound transferrin at the plasma membrane
(reviewed by Hansen et al., 1993; Alberts et al., 1994). Following endocytosis, iron


123


88
+ 10% FBS for 1 h at 37C before fixing with -20C MeOH and staining with anti-lpgl20
antibody (Figure 4-13 A-D). The normal fibroblasts, without chloroquine treatment,
contained small lysosomes distributed throughout the cytosol, as visualized by labeling
with the lpgl 20 antibody (Figure 4-13 A). However, following chloroquine treatment,
the lpgl20-containing compartment in the normal fibroblasts (Figure 4-13 B) increased in
size and number resembling the lysosomal-like vacuoles of the untreated LPI cells
(Figure 4-13 C). When the LPI fibroblasts were treated with chloroquine, the lpg 120-
containing vacuoles grew to an enormous size and clustered tightly in the perinuclear
region of the cells (Figure 4-13 D). Incubating both cell types with a higher
concentration of chloroquine, or for an extended period of time resulted in similar, yet
even more pronounced effects. The vacuoles in the normal fibroblasts became even
larger and more abundant, whereas the intracellular membranes of the LPI cells appeared
to dissolve until only a few giant vacuoles filled the entire intracellular space (data not
shown).
Antibody labeling of normal and LPI fibroblasts expressing GFP-CAT1. The
experiments discussed in this section are an extension of those presented above. As
mentioned before, no differences were detected in the abundance or distribution of
endogenous or exogenous CAT1 in normal and LPI fibroblasts. Of all of the organelle-
specific antibodies tested, only those associated with the lysosomes showed any
abnormalities in the LPI cells. The following co-localization experiments examined
whether or not the CAT1 transporter, in LPI cells, was trapped in a particular
compartment or vesicle pool, such as the lysosomes. Normal and LPI fibroblasts were
transfected with the GFP-CAT1 fusion protein according to the lipofectamine protocol


features between the two is the acceptance of glutamine by ASCT2, but not by ASCT1
(Utsunomiya et al., 1996).
ASCT1 exhibits a unique pH-dependent substrate specificity (Tamarappoo et al.,
1996), first described for System ASC by Christensen and colleagues (Vadgama and
Christensen, 1984). At neutral pH, ASCT1 preferentially transports zwitterionic amino
acids, whereas if the assay pH is lowered to 5.5, the transporter will accept both
zwitterionic and anionic amino acids. Transport assays using anionic amino acid analogs
to compete with the zwitterionic substrates at pH 5.5, indicate that the substrates may
share the same binding region (Tamarappoo et al., 1996). One hypothesis to explain this
pH effect is that one or more of the eight histidine residues in ASCT1 accept(s) a positive
charge when the pH is lowered below 6.0 (the pKa for the histidine side chain). The
positively charged histidine(s) may serve as a binding site for negatively charged amino
acids, such as glutamate, aspartate, and cysteate. ASCT2 also has a low affinity for
glutamate at neutral pH. and the affinity for the anionic amino acid increases as the assay
pH is lowered (Utsunomiya et al., 1996; Kekuda et al., 1996).
Trafficking of Membrane Proteins
The cloning and expression of a number of the amino acid transporters have led to
the generation of sequence-specific antibodies from corresponding peptides and fusion
proteins. The availability of antibodies and the development of high-resolution
microscopes have allowed investigators to initiate studies on the cell biology of amino
acid transporters, an area of the field in which little is known. The first mammalian
amino acid transporter antibody was produced in 1992. In recent years, many of the


Figure 4-13. Lysosomal detection in normal and LPI fibroblasts following chloroquine
treatment. Normal (A and B) and LPI (C and D) fibroblasts were treated with
chloroquine and stained with anti-lpgl20 antibody according to the protocol described in
Chapter 4. Briefly, cells were incubated in MEM + 10% FBS (A and C) or MEM + 10%
FBS containing 50 mM chloroquine (B and D) for 1 h before fixing with -20C MeOH
and staining with anti-lpgl20 antibody. The anti-lpgl20 primary antibody was detected
with a goat anti-rabbit IgG linked to FITC. Staining from three independent experiments
was analyzed by deconvolution microscopy and shown to be reproducible. The data
shown represent analysis of 0.2 pm sections through the cells.


24
protocol (InVitrogen). A partial digest was performed using EcoRI restriction enzyme to
isolate the EAAT1 1680 base pair fragment. This fragment was then subcloned into the
pEGFP(N3) vector, at the EcoRI site, in order to place the EAAT1 cDNA before the
GFP. The 1680 base pair fragment was also cloned into the EcoRI site of the pEGFP(C3)
vector, which placed EAAT1 just before a stop codon at the C-terminus of the GFP.
A 2280 base pair fragment, including the entire coding sequence and stop codon,
of the CAT1 cDNA was cut out of pCDNA3 with Hindlll and EcoRI and subcloned
directly into the multiple cloning site of pEGFP(C3) at the Hindlll and EcoRI sites, thus
placing CAT1 at the C-terminal end of GFP.
Mutagenesis. Oligonucleotide-directed site-specific mutagenesis of the CAT1
cDNA was performed using the MORPH Mutagenesis kit from 5 Prime 3 Prime, Inc.
(Boulder, CO). CATMUT1 was constructed using the wild type GFP(C3)-CAT1 cDNA
as the template and the CAT1.1 mutant oligonucleotide, GGT GTT GAG GGA GCG
GGA CAG GCG GCT CTC CTC CCG GCT GGA GTC GAC CAC CTT CCG GCG.
This mutagenesis reaction resulted in the substitution of serine for cysteine at residues 20
and 30. CATMUT2 was constructed using the GFP(C3)-CAT1 cDNA as the template
and the CAT1.2 mutant oligonucleotide, (GCA TCT GCT GGC CCA GCC CGA GCA
GGT TTT TGG AGGCCA TTG TGC TGA GCG AAT CTG C). This reaction resulted
in the substitution of alanine for glycine at position 2. CATMUT3 was generated using
CATMUT1 as the template and the mutant oligonucleotide, CAT1.3 (GCA TCT GCT
GGC CCA GCC CGA GCA GGT TTT TGC AGG CCA TTG TGC TGA GCG ATT
CTG C). This reaction resulted in the substitution of alanine for glycine at position 2,


46
the majority of the P-integrin labeling appeared in the plasma membrane-enriched
fraction, the band in the nuclear fraction suggests that the 300 x g fraction is
contaminated with plasma membrane, or more likely unbroken cells. When a 1:100
dilution of EAAT1-C was incubated with PAEC fractions, a band at approximately 75
kDa was detected in all three fractions in almost equal amounts (data not shown).
Transfection of EAAT1-GFP constructs in human fibroblasts. Two fusion
constructs between EAAT1 and green fluorescent protein (GFP) were generated to
compare the localization of the expressed transporter with the endogenous EAAT1
immunofluorescence experiments. The EAAT1-GFP(N3) fusion protein was constructed
with the GFP tag at the C-terminal end of EAAT1. and the GFP(C3)-EAAT1 was
constructed with GFP at the N-terminal end. The GFP tag was attached to either the N-
or C-terminus of EAAT1 to ensure that the location of the GFP did not interfere with
normal transporter trafficking (see Methods Chapter for details). Both of these constructs
were expressed in human fibroblasts using the lipofectamine transfection procedure
described in the Methods Chapter. After 24 hours of expression, cells were fixed with
-20C MeOH. mounted on glass slides with Fluoromount-G, and visualized using an
FITC filter on the deconvolution microscope, fluman fibroblasts transfected with the
GFP vector only (Figure 3-7 A) showed diffuse fluorescence throughout the entire cell,
whereas cells transfected with either EAAT1-GFP(N3) (Figure 3-7 B) or GFP(C3)-
EAAT1 (Figure 3-7 C) showed distinct vesicles in the cytoplasm as well as staining of
the plasma membrane. After analysis of three separate experiments, it was concluded that
there was no difference in the fluorescent patterns generated by the two different GFP


153
4.0 urn
4.0 uin


178
Kanai, Y. and Hediger, M. A. (1992) Nature 360, 467-471
Kanai, Y., Nussberger, S., Romero, M. F., Boron, W. F., Hebert, S. C. and Hediger, M.
A., (1995) J. Biol. Chem. 270, 16561-16568
Kanai, Y., Stelzner, M. G., Lee, W.-S., Wells, R. G., Brown, D. and Hediger, M. A.
(1992) Am. J. Physiol. 263, F1087-F1093
Kekuda, D. K., Prasad, P. D., Fei, Y.-J., Torres-Zamorano, V., Sinha, S., Yang-Feng, T.
L., Leibach, F. H. and Ganapathy, V. (1996) J. Biol. Chem. 271, 18657-18661
Kawakami, H., Tanaka, K., Nakayama, T., Inoue, K. and Nakamura, S. (1994) Biochem.
Biophys. Res. Commun. 199, 171-176
Kerem, B.-S., Rommens, J. M., Buchanan, J. A., Markiewicz, D., Cox, T. K.,
Chakravarti, A., Buchwald, M. and Tsui, L.-C. (1989) Science 245, 1073-1080
Kilberg, M. S., Stevens, B. R and Novak, D. (1993) Annu. Rev. Nutr. 13, 137-165
Kim, J. W., Class, E. I., Albritton, L. M., and Cunningham, J. M. (1991) Nature 352,
725-728
Klockner, U Storck, T., Conradt, M. and Stoffel, W. (1993) J. Biol. Chem. 268, 14594-
14596
Komfeld, S. and Mellman, I. (1989) Annu. Rev. Cell Biol. 5, 483-525
Kurz, S. and Harrison, D. G. (1997) J. Clin. Invest. 99, 369-370
Laemmli, U. K. (1970) Nature 227, 680-685
Laine, R. O. and Esser, A. F. (1989) J. Immunol. 143, 553-557
Li, S., Mallory, M., Alford, M., Tanaka, S. and Masliah, E. (1997) J. Neuropathol. Exp.
Neurol. 56,901-911
Liao, K. and Lane M. D. (1995) Biochem. Biophys. Res. Commun. 208, 1008-1015
Lin, C. L., Bristol, L. A., Jin, L., Dykes-Hoberg, M., Crawford, T., Clawson, L. and
Rothstein, J. D. (1998) Neuron 20, 589-602
Lisanti, M. P., Scherer, P. E., Tang, Z., and Sargiacomo, M. (1994) Trends Cell Biol. 4,
231-235
Liu, J., Garcia-Cardena, G., and Sessa, W. C. (1996) Biochemistry 35, 13277-13281
Ljungdahl, P. O., Gimeno, C. J., Styles, C. A. and Fink, G. R. (1992) Cell 71, 463-478


127
et al., 1997). The caveolin scaffolding domain, which includes amino acids 82-101, has
been reported to interact with Ga subunits, Ha-Ras, Src family tyrosine kinases, and
eNOS (Song et al., 1996; reviewed by Okamoto et ah, 1998). With the exception of Ha-
Ras, a peptide encoding the scaffolding domain sequence can functionally inactivate each
of these proteins (Song et ah, 1996). The bradykinin and p-adrenergic receptors are
dependent on ligand binding for migration to caveolae, whereas the caveolar localization
of the endothelin receptor and amyloid precursor protein is ligand-independent (Chun et
ah, 1994; Feron et ah, 1997; de Weerd et ah, 1997).
eNOS has been shown to localize to the Golgi in cultured bovine aortic
endothelial cells (EC), human umbilical vein EC, and intact human blood vessels (Morin
and Stanboli, 1993; O'Brien et ah, 1995; Sessa et ah, 1995). In 1996, several laboratories
also documented significant localization of eNOS to plasmalemma caveolae of cultured
bovine lung microvascular EC and in luminal regions of plasmalemma isolated from
intact, perfused rat lungs (Garcia-Cardena et ah, 1996; Shaul et ah, 1996). For example,
Garcia-Cardena et ah showed that caveolin and membrane-bound eNOS co-localize in
lung microvascular endothelial cells and antibodies against one could be used to
immunoprecipitate the other, strongly suggesting that eNOS is complexed with caveolin
(Garcia-Cardena et ah, 1996). More recently, several labs have reported a predicted
caveolin-binding sequence (FSAAPFSGW) in the catalytic domain of eNOS that
interacts with caveolin-1 at residues 82-101. It is known that the aromatic residues are
essential for the recognition of caveolin-1, and when Sessa and colleagues mutated the
aromatic amino acids to alanine, caveolin no longer inhibited eNOS activity (Garcia-


8
Glutamate transport studies in salamander retina glial cells demonstrated that glutamate
uptake is electrogenic and coupled to the co-transport of three Na* ions and a FT, as well
as the counter-transport of one K ion (Zerangue and Kavanaugh, 1996). In the central
nervous system, the stoichiometry of the glutamate and ion transport must be tightly
regulated. Ischemic conditions following a stroke may lead to the breakdown of
electrochemical gradients as a result of lower ATP levels and reduced functioning of
Na\K ATPase proteins (Szatkowski and Attwell, 1994). If the ion gradients are
disrupted, then it is believed that the glutamate transporters can function in reverse,
resulting in the release of glutamate into the synaptic cleft, and subsequent neurotoxicity
and neuronal death (Kanai et al., 1995).
In an attempt to isolate galactosyltransferase from rat brain, Storck and coworkers
co-purified the first glutamate transporter, designated GLAST1 (Storck et al., 1992). The
isolated protein showed homology to the previously cloned bacterial glutamate and
monocarboxylate transporters, and its function as a glutamate transporter was confirmed
by expression in Xenopus oocytes followed by radiolabeled amino acid uptake (Klockner
et al., 1993). Pines and coworkers cloned the second glutamate transporter, GLT1, by
screening a rat cDNA expression library with antibodies generated against the partially
purified protein (Pines et al., 1992). Northern analysis and in situ hybridization detects
GLAST1 and GLT1 mRNA expression primarily in glial cells of the central nervous
system (Storck et al., 1992; Otori et al., 1994), where they play an important role in
clearing toxic levels of glutamate from the synaptic clefts. Although the mechanism is
unclear, recent data from Rothstein and coworkers indicate that abnormal GLT1 mRNA
species may be responsible for the decreased glutamate transport detected in patients with


18
MCAT1 cDNA was a generous gift from Dr. James Cunningham at Brigham and
Women's Hospital (Boston. MA), and the rat EAAC1 cDNA was cloned from a rat
hippocampal librar)' (Velaz-Faircloth et al., 1996). All of the antibodies are described in
the chapters in which they were used.
Methods
Cell culture. Pulmonary artery endothelial cells (PAEC) were prepared by
collaborators in Dr. Edward Block's laboratory. PAEC were isolated by collagenase
treatment of the main pulmonary artery of 6-7-month-old pigs and were cultured for 3-7
passages as described by Block and coworkers (Block et al., 1989). One hundred-mm
dishes or 6-well trays were incubated with 7-10 pg/ml of fibronectin (Sigma Chemical
Co.), dissolved in RPMI medium, overnight at 37C under a humidified atmosphere of
5% CO:-95% air. Prior to plating cells, the fibronectin solution was aspirated, and
dishes/trays were allowed to dry for 30 min in a culture hood under a UV lamp. Once
plated, PAEC were maintained in RPMI + 4% or 10% fetal bovine serum (FBS). Hela,
HepG2 (human hepatoma), HEK 293 (human embryonic kidney), CHO and BNL.CL2
(mouse hepatocytes) cells were maintained in Eagle's minimal medium (MEM) + 4%
FBS. Cultured fibroblasts from Finnish LPI patients and sex-age-matched normal
controls were obtained from Dr. Olli Simell at the Central Hospital, University of Turku
(Turku, Finland). The fibroblasts were cultured in 75-mm flasks and maintained in
MEM. supplemented with 10% FBS. The cells were passaged after achieving
approximately 80% confluence, and were used for experiments until the sixth to eighth


149


CHAPTER 5
CAVEOLAR COMPLEX BETWEEN THE CATIONIC AMINO ACID
TRANSPORTER 1 AND ENDOTHELIAL NITRIC OXIDE SYNTHASE
Introduction
The nitric oxide synthases (eNOS, iNOS, and nNOS) catalyze the conversion of
L-arginine to L-citrulline and nitric oxide (NO), a nitrogen-centered free radical with
multiple and unique physiologic and bioregulatory activities (reviewed by Moneada and
Higgs, 1993; Nathan and Xie, 1994). Pulmonary endothelial cells generate NO via the
catalytic action of an NADPH-requiring, Ca+7calmodulin-dependent NO synthase
(eNOS) that is membrane-associated (Moneada et al., 1991; Forstermann et ah, 1994;
Palmer et ah, 1988). Pulmonary endothelial cells are a rich source of nitric oxide (NO)
which functions as both a paracrine and autocrine mediator. As a paracrine mediator, NO
controls vascular smooth muscle tone and inhibits leukocyte adhesion and platelet
aggregation. As an autocrine mediator, NO influences growth factor signals and cellular
proliferation. Many cardiovascular diseases, including hypertension, diabetes,
atherosclerosis, and heart failure, develop complications because endothelial cells fail to
produce adequate amounts of NO.
In endothelial cells, eNOS-mediated formation of NO from arginine is dependent
upon an adequate and continuing supply of arginine (Aisaka et ah, 1989; Cooke et ah,
1991; Taylor and Poston, 1994; Rossitch et ah, 1991). Several studies have shown that
the half-saturating arginine concentration for eNOS is less than 10 pM (Liu et ah, 1996;
124


Figure 5-9. Immunofluorescent staining of PAEC transfected with the CATMUT1
palmitoylation mutant. PAEC were transfected for 3 h with the CATMUT1 fusion
protein that has serine substitutions for cysteine at residues 20 and 30 and was prepared
according to the mutagenesis protocols described in the Methods Chapter. Transfection
was performed as outlined in the Methods Chapter. Following 24-48 h of expression,
cells were fixed with -20C MeOH and stained with antibodies against caveolin (A) and
eNOS (B). Both the caveolin and eNOS antibodies were detected with a goat anti-rabbit
IgG conjugated to Texas Red and visualized by deconvolution microscopy. Images were
processed from three independent experiments and the staining was determined to be
reproducible. The data shown represent analysis of 1.0 pm sections through the cells.


human fibroblasts, whereas the EAAT3 transporter was observed in intracellular vesicle
compartments and in concentrated clusters on the plasma membrane. Lysinuric Protein
Intolerance (LPI) provided a model system for investigating the subcellular organization
and trafficking pathways in a disease with visual morphological defects. A survey of
organelle integrity led to the discovery of an abnormal population of lysosomes in the
LPI fibroblasts. Future studies will investigate the contents of the LPI lysosomes, as well
as the amino acid transport across lysosome-enriched membrane preparations.
Immunostaining of porcine aortic endothelial cells (PAEC) revealed "patches" of the
cationic amino acid transporter CAT1 that co-localized with antibodies against caveolin
and endothelial nitric oxide synthase (eNOS). When incubated with solubilized PAEC
plasma membrane proteins, an eNOS-specific antibody immunoprecipitated CAT1-
specific arginine transport activity. These results document the existence of a caveolar
complex between CAT1 and eNOS in PAEC that provides a potential mechanism for the
efficient delivery of the arginine substrate to eNOS for nitric oxide (NO) production. The
individual projects presented in this thesis share a common goal in documenting the
cellular localization, and when possible, understanding the functional consequences, of
the transporter protein distribution.
x


34
aspartate transport across epithelial cells of the kidney and intestine. In some cases,
mental retardation or neurological abnormalities are also symptoms of the disease,
however, there was no evidence of neurological damage in the EAAT3 knockout mice.
The dicarboxylic aminoacidurea acquired by the knockout mice is explained by the fact
that EAAT3 is strongly expressed in the kidney and is responsible for tubular
reabsorption of glutamate and aspartate from the glomerular filtrate.
Based on their accepted function of Na'-dependent glutamate/aspartate transport
and the membrane spanning structure of these amino acid transporters, we predicted that
the EAAT transporters would be primarily localized to the plasma membrane, with some
intracellular pools involved in biosynthesis or recycling. However, this could not be
assumed based on the recent detection of several plasma membrane proteins in the
nuclear membrane and matrix. P-glycoprotein. a 170 kDa protein with 12 membrane
spanning domains, has been implicated in conferring multi-drug resistance in cancer cells
(Juliano and Ling. 1976). It accomplishes this by actively exporting a wide-range of
chemotherapeutic agents out of the cell in an ATP-dependent mechanism. Recently. P-
glycoprotein was reported to reside in the nuclear membrane and matrix as well as on the
plasma membrane (Baldini. 1995). Also, a number of laboratories have reported the
presence of different growth factor receptors associated with the nucleus (Stachowiak et
al., 1996; Podlecki et ai 1987; Rakowicz-Szulczvnska et al., 1989). Traditionally, these
receptors were believed to reside on the plasma membrane and transmit a signal from the
extracellular environment to the cytosol upon ligand binding. However, the fibroblast,
insulin, and epidermal growth factor receptors are all integral membrane proteins that
have been localized to the nucleus.


81
IgG-purified, and affinity-purified antibodies were used at dilutions ranging from 1:25 to
1:1000. In addition, cells were fixed with either -20C MeOH or 2%-4%
paraformaldehyde, for 10 to 30 minutes, before permeabilizing and incubating with the
CAT1 antibody. Several different normal and LPI patient cell lines were used and
antibody incubations ranged from 1 h to overnight (data not shown). However, none of
these conditions produced the CATI-containing LPI vesicles seen previously. Additional
research, perhaps with new antibody preparations, will be required to confirm the CAT1
staining of LPI fibroblasts.
Previous LPI research in the Kilberg laboratory was limited to the study of the
CAT1 transporter, yet it is entirely possible that the proposed trafficking defect involves
other transporter, and/or non-transporter, proteins. Antibodies against several other
nutrient transporters were available and used for intra- and extracellular
immunofluorescent labeling of normal and LPI fibroblasts. Both cell types were stained
with a 1:50 dilution of the NaMndependent glucose transporter, GLUT1, a 1:200 dilution
of EAAT3, and a 1:50 dilution of EAAT1 following both -20C MeOH and 4%
paraformaldehyde fixation. Each of these antibodies was detected using a goat anti-rabbit
IgG conjugated to FITC. GLUT1 labeling was detected in small amounts on the plasma
membrane, but primarily in cytoplasmic and perinuclear vesicles. The EAAT3 antibody
was detected in clusters on the plasma membrane, as well as in vesicles throughout the
cytoplasm. The EAAT1 antibody, discussed in detail in Chapter 3, was localized to the
nuclear membrane. No differences in the staining patterns were detected, and no
abnormal intracellular vesicle populations were identified, when labeling the normal and
LPI cells with antibodies against any of these transporters (data not shown).


66


25
and serine for cysteine at positions 3, 20, and 30. Each mutagenic oligonucleotide above
was designed with an internal restriction site for analyzing the success of the mutagenesis
reaction. A Sal I restriction site was engineered into the CAT 1.1 mutagenic
oligonucleotide and the CAT1.2 and CAT1.3 mutagenic oligonucleotides were
constructed with an internal Ava I site. Prior to beginning the mutagenesis, 2.5 pg of the
oligonucleotide was 5' phosphorylated in a reaction using 5 pi of 10X T4 polynucleotide
kinase buffer, 25 U of T4 polynucleotide kinase, and 10 mM ATP. This reaction was
allowed to proceed at 37C for 1 h before being terminated by heating to 65C for 10 min.
The annealing procedure involved mixing 0.03 pmol of the target cDNA (GFP-CAT1),
2.0 pi of 10X MORPH annealing buffer, and 100 ng of phosphorylated mutagenic
oligonucleotide and heating the solution to 100C for 5 min to denature the double-
stranded DNA. At this point, solutions were either placed at room temperature for 30
minutes, or placed in a beaker of 70C water that was allowed to cool slowly to room
temperature. One procedure worked better than the other for specific oligonucleotides
and the choice was determined experimentally. For the replacement strand reaction, 8 pi
of 3.75X MORPH synthesis buffer, 3 U T4 DNA polymerase, and 4 U T4 DNA ligase
were added directly to the annealing reaction and incubated for 2 hr at 37C, then for 15
min at 85C to terminate the reaction. A 1:10 dilution of Dpn I restriction enzyme was
prepared and 1 pi was added to each mutagenesis reaction. The solution was incubated
for 30 min at 37C and then placed on ice for 5 min. The premise behind this digestion is
that Dpn I will specifically digest only double-stranded DNA in which both strands are
methylated, therefore, any double-stranded non-mutagenized target plasmid DNA will be


36
independent sources. The EAAT1 cDNA was used to create a fusion protein with the
green fluorescent protein (GFP) and distribution of the transporter was examined
following transient expression of this EAAT1-GFP fusion protein. These experiments
were performed in combination with immunofluorescence and immunoblotting using four
different antibodies specific to EAAT1.
Methods
Glutamate transporter antibody production. The polyclonal EAAT3 glutamate
transporter antibody was raised in rabbit against an EAAT3-maltose binding protein that
was constructed by Dr. Marc Malandro using the C-terminal 120 amino acids of the rat
EAAT3 (EAAC1) sequence. The glutamate transporter antibody designated EAAT1-R
was obtained from Dr. Jeffrey Rothstein at Johns Hopkins University. EAAT1-R is a
polyclonal antibody that was raised in rabbit against the amino acid residues 3-17
(KSNGEEPRMGSRMGR) at the N-terminus of the human EAAT1 protein (Ginsberg et
al., 1995). The EAAT1 glutamate transporter antibody designated EAAT1-S was
obtained from Dr. Wilhelm Stoffel at the University of Cologne (Cologne, Germany).
This polyclonal antibody was generated in rabbit against a peptide consisting of amino
acid residues 24-40 (KRTLLAKKKVQNITKED) at the N-terminus of the rat EAAT1
sequence (Wahle and Stoffel, 1996). The EAAT1-C polyclonal antibody was generated
in guinea pig, by Chemicon International, Inc. (Temecula, CA), against the rat C-terminal
peptide, QLIAQDNEPEKPVADSETKM (Storck et al., 1992). The polyclonal EAAT1-
D antibody was purchased from a-Diagnostics International (San Antonio, TX). This
antibody was raised in rabbit against a peptide (amino acids 504-518,


14
molecules are released from the transferrin in response to the low pH of the endosme.
This allows the unbound transferrin to recycle to the plasma membrane with its receptor.
The transferrin molecule is then freed so that it may sequester more extracellular iron (De
Silva et al., 1996).
The final trafficking pathway that contributes to the life cycle of membrane
proteins involves the transport of proteins to the lysosomes for degradation (reviewed by
Komfeld and Mellman, 1989), or the ubiquitination and subsequent degradation by
proteosomes (reviewed by Hershko and Ciechanover, 1992). Materials from multiple
pathways are emptied into the lysosomes, where acid hydrolases function in the regulated
digestion of macromolecules such as membrane proteins. The mannose-6-phosphate
receptor (M6PR) binds to the phosphorylated mannose modification on lysosomal-
destined digestive enzymes and shuttles these acid hydrolases from the TGN to the late
endosomes. The acid hydrolases eventually end up in lysosomes and the M6PR recycles
to the TGN. Therefore, the M6PR is an excellent marker for the late endosomal and TGN
compartments. A second trafficking pathway that leads to degradation evolves from
endocytosis. In a process that is poorly understood, materials destined for degradation
are transferred from early endosomes to late endosomes to lysosomes (Alberts et al.,
1994). Antibodies against mammalian amino acid transporters have become available
only in the last couple of years, so no information has been published concerning the
molecular mechanisms by which these transporters are degraded.


174
Baldini, N., Scotlandi, K., Serra, M., Shikita, T., Zini, N., Ognibene, A., Santi, S.,
Ferracini, R. and Maraldi, N. (1995) Eur. J. Cell Biol. 68, 226-239
Bar-Peled, O., Ben-Hur, H., Biegon, A., Groner, Y., Dewhurst, S., Furuta, A. and
Rothstein, J. D. (\997) J. of Neurochem. 69,2571-2579
Baydoun, A. R., Emery, P. W., Pearson, J. D., and Mann, G. E. (1990) Biochem.
Biophys. Res. Commun. 173, 940-948
Bertrn, J., Magagnin, S., Werner, A., Markovich, D., Biber, J., Testar, X., Zorzano,
Kuhn, L. C., Palacin, M. and Murer, H. (1992) Proc. Nat. Acad Sci. U.S.A. 89,
5606-5610
Bhat, G. B., and Block, E. R. (1992) Am. J. Respir. Cell. Mol. Biol. 6, 633-638
Bhat,G. B. and Block, E. R. (1990) Am. J. Respir. Cell. Mol. Biol. 3, 363-367
Bliss, T. V. and Collingridge, G. L. (1993) Nature (Lond.) 361, 31-39
Block, E. R., Herrera, H., and Couch, M. (1995) Am. J. Physiol. 269, L574-L580
Block, E. R Patel, J. M. and Edwards, D. (1989) Cell Physiol. 26, C223-C231
Boger, R1 H., Bode-Boger, S. M., Thiele, W., Junker, W., Alexander, K. and Frolich, J.
C. (1997) Circulation 95, 2068-2074
Boulikas, Teni (1996) J. Cell Biochem. 60,61-82
Bouvier, M., Szatkowski, M., Amato, A. and Attwell, D. (1992) Nature 360, 471-474
Bu, G. and Schwartz, A. L. (1994) in Cell Biology: A Laboratory Handbook pp. 199-
200, Academic Press, Inc.
Buchanan, J. A., Rosenblatt, D. S. and Scriver, C. R. (1985),4wj. NY. Acad. Sci. 256,
401-402
Bussolati, O., Laris, P. C., Nucci, F. A., Dall'Asta, V., Longo, N., Guidotti, G. G. and
Gazzola, G. C. (1987) Am. J. Physiol. 253, C1-C7
Bussolati, O., Sala, R., Astorri, A., Rotoli, B. M., DallAsta, V., and Gazzola, G. C.
(1993) Am. J. Physiol. 265, C1006-C1014
Calonge, M. J., Gasparini, P., Chillaron, J., Chilln, M., Gallucci, M., Rousaud, F.,
Zelante, L., Testar, X., Dallapiccola, B., Di Silverio, F., Barcelo, P., Zorzano, A.,
Nunes, V. and Palacin, M. (1994) Nat. Genet. 6, 420-425
Carlsson, M. and Carlsson, A., (1990) Trends Neurosci. 13, 272-276


82
Localization of expressed CAT1 in normal and LPI fibroblasts. The experiments
investigating endogenous CAT1 localization relied on the detection of the human
transporter with a anti-murine CAT1 antibody. An expression system was developed in
order to confirm endogenous experiments, as well as, to avoid any technical problems
associated with the CAT1 antibody. Transfection with several CAT1 constructs was
attempted before deciding to use the green fluorescent protein (GFP) vector from
Clontech (Palo Alto, CA). These constructs included a FLAG-tagged murine CAT1
cDNA, generated in our laboratory, as well as an HA-tagged human CAT1 cDNA and a
human CAT1-GFP fusion protein, both provided by Dr. Lorraine Albritton at University
of Tennessee Medical Center. Neither the HA-tagged CAT1 nor the FLAG-tagged CAT1
constructs were detected following transfection, and the expression of the CAT 1-GFP
from Dr. Lorraine Albritton was extremely low (data not shown). Dr. Lorraine Albritton's
CAT 1-GFP fusion protein was constructed with a GFP variant that was not modified for
enhanced expression and fluorescence in human cells. Therefore, a GFP(C3)-CAT1
fusion protein was constructed using Clontechs humanized GFP variant and the CAT1
cDNA sequence from mouse (See Methods Chapter).
The GFP tag (described in the Methods Section) provided an alternative technique
for observing the localization of the CAT1 protein in the normal and LPI fibroblasts. If
the CAT1 transporters were trafficking incorrectly in LPI cells, but the available antibody
could not detect it, then the GFP(C3)-CAT1 would provide an independent method of
CAT1 localization. Transfection of normal fibroblasts with the GFP(C3) only, which
lacks a targeting signal and, therefore, diffuses throughout the cytoplasm and nucleus,
was used as a control (Figure 4-4 A). Alternatively, transfection of normal fibroblasts


140
evidence for a direct interaction between CAT1 and eNOS, CAT1 antibody-coated
agarose beads were generated as described above for binding of anti-eNOS to the beads
and used to immunoprecipitate CAT1 from detergent solubilized PAEC plasma
membrane vesicles. As controls, non-coated or non-immune IgG-coated agarose beads
were used in the immunoprecipitation. Proteins recovered from pelleted beads, as well as
the supernatant, were run on a SDS-PAGE gel and immunoblotted with an anti-eNOS
antibody (Figure 5-8). eNOS was detected in both the supernatants from control
immunoprecipitations as well as immunoprecipitations with the CAT1 antibody, although
less protein appeared in the latter supernatant. When the immunoprecipitates were
immunoblotted with anti-eNOS antibody, eNOS protein was detected primarily in the
control immunoprecipitation presumably the result of non-specific adsorption, with little
protein in the sample immunoprecipitated with the CAT1 antibody. This experiment was
repeated five times in order to optimize the conditions and refine the technique, however,
each attempt at co-precipitating eNOS with the anti-CATl antibody was unsuccessful.
Covalently cross-linking by treating PAEC membrane vesicles with dimethyl
suberimidate (DMS) prior to immunoprecipitation with the anti-CATl antibody was
performed. Chemical cross-linking reagents introduce covalent bonds between
neighboring proteins that are tightly associated. DMS, specifically, is one of a group of
chemicals that reacts with the e-amino groups of lysines and available N-terminal amines
(Bu and Schwartz, 1994). The cross-linking was attempted in the event that the CAT1-
eNOS interaction was being disrupted during the immunoprecipitation procedure. The
same immunoprecipitation procedure was followed (see Methods Section of this chapter),


6
mediates the high-affinity, Na*-independent transport of cationic amino acids and shares
the greatest homology with the CAT1 family member. In the same year, the mouse
CAT3 cDNA was identified by Ito and Groudine during an attempt to isolate germ-layer
specific transcripts from mouse embryos (Ito and Groudine, 1997). The mouse CATS
was localized to the brain by in situ hybridization studies, and exhibited the same
structural and transport characteristics as the rat homolog. The highly conserved tissue-
specificity between the rat and mouse proteins suggests an important role for CAT3 in the
brain. Detection of mRNA in brain capillaries suggests a role for CAT3 in the transport
of cationic amino acids by endothelial cells at the blood-brain barrier (Hosokawa et al.,
1997). Just as CAT2 may provide the inducible nitric oxide synthase (iNOS) with
substrate for NO production in T cells, it is feasible that CAT3 provides the neuronal
nitric oxide synthase (nNOS) isoform with the arginine required for NO production in the
nervous system.
The CAT family, comprising four functionally similar yet distinct proteins, is
only one of two families responsible for the transport of cationic amino acids. The other
family has been identified based on mRNA expression in Xenopus oocytes and currently
includes two proteins, NBAT and 4F2hc. Expression of either in oocytes induces Na"-
independent transport of cationic amino acids, but NBAT mediates the uptake of Na+-
independent zwitterionic amino acids whereas 4F2hc-catalyzed uptake of these substrates
is Na-dependent (Bertrn et al., 1992; Wells et al., 1992). These specificities correspond
to two known transport activities, called System b+ (NBAT) and System y+L (4F2hc),
respectively. The hydropathy plots of both transporters predict either one or four
membrane spanning domains. This is an interesting feature for two proteins that are


[
SUBCELLULAR LOCALIZATION AND TRAFFICKING
OF AMINO ACID TRANSPORTERS
By
KELLY KRISTIN MCDONALD
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1998


Figure 4-2. Morphology of normal and LPI fibroblasts by electron microscopy. Normal
and LPI fibroblasts were fixed in a solution of 4% paraformaldehyde and 0.5%
glutaraldehyde for 30 min before pelleting. The cells were then dehydrated in ethanol,
embedded in an acrylic resin (either K4M or unicryl), and thin-sectioned for mounting on
nickel grids according to the protocol of the Electron Microscopy Core (University of
Florida). Electron micrographs were taken of normal fibroblasts at a magnification of 6K
(A), LPI fibroblasts at 6K (B), LPI fibroblasts at 9K (C), and LPI fibroblasts at 25K (D).


170
The experiments performed in Chapter 3 were designed to compare and contrast
the cellular localization of amino acid transporters within the same gene family.
Although antibodies to four of the five cloned glutamate transporters were available, only
EAAT1 and EAAT3 were detected in human fibroblasts by immunofluorescence.
Whereas both of the glutamate transporters were detected in intracellular vesicle pools,
additional immunoreactivity for EAAT3 was shown to cluster on the plasma membrane,
whereas that for EAAT1 was observed in the nuclear membrane. Both of these
observations were interesting and unexpected. Although EAAT1 and EAAT3 belong to
the same gene family, the EAAT3 clustering pattern is similar to that of the unrelated
CAT1 transporter. The validity of the EAAT1 nuclear staining remains in question;
however, it raises some intriguing possibilities. Even if the EAAT1-R antibody is
detecting a protein other than EAAT1, it is possible that a protein related to EAAT1
resides in the nucleus. More extensive immunoblotting must be performed to determine
if contamination of other membranes was contributing to the immunoreactivity of
EAAT1 in the nuclear fraction. Conclusive proof must await antibody-independent
identification of the immunoreactive protein within the nuclear membrane.
The experiments in Chapter 4 were designed to test the hypothesis that the
abnormal amino acid transport observed in LPI fibroblasts was due to aberrant trafficking
of membrane proteins, including transporters. Although the CAT1 staining appeared to
be normal, a thorough investigation of the organelles involved in the cellular trafficking
pathways detected a morphological difference between the lysosomes of the normal and
LPI fibroblasts. It is suspected that the enlarged and overly abundant LPI vacuoles are
lysosomes based on their detection by antibodies against cathepsin D and the lysosomal


Figure 4-6. Intracellular staining of normal and LPI fibroblasts with antibodies against
proteins of the nuclear membrane and nucleolus. According to the protocol described in
the Methods Chapter, normal (A and B) and LPI (C and D) fibroblasts were fixed with
-20C MeOH and subjected to immunohistochemistry with the anti-414 antibody specific
for the nuclear membrane (A and C) and the anti-D77 antibody specific for the nucleolus
(B and D). Both primary antibodies were visualized with a goat anti-mouse IgG
conjugated to FITC. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. One cell in each panel is
outlined to distinguish the cell periphery. The data shown represent analysis of 0.2 pm
sections through the cells.


Figure 5-4. Detection of CAT1 and eNOS in the Golgi of PAEC. According to the
immunofluorescence protocol described in the Methods Chapter, PAEC were fixed with
-20C MeOH and co-stained with antibodies against CAT1 and 10e6 (A) or eNOS and
sialyltransferase (ST) (B). The anti-10e6 and anti-ST are antibodies specific for proteins
of the Golgi Complex. The anti-CAT 1 and anti-ST were detected with goat anti-rabbit
IgG linked to FITC, whereas anti-eNOS and anti-10e6 were labeled with goat anti-mouse
IgG linked to Texas Red. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells.


33
concentrations as well as neurodegeneration and progressive paralysis (Rothstein et al..
1996). Antisense oligonucleotides to EAAT1. EAAT2. and EAAT3 were administered
intraventricularly to male Sprague-Dawley rats for 7-10 days. Within 3 days, the animals
that were treated with the EAAT1 and EAAT2 antisense oligonucleotides began to show
evidence of progressive motor degeneration, and by day 8. their hindlegs were paralyzed.
When extracellular levels of glutamate were measured with microdialvsis probes in the
ipsilateral striatum of treated rats, extracellular glutamate concentrations were elevated by
32-fold in EAAT2 antisense rats, and by 13-fold in EAAT1 antisense rats. Loss of
EAAT3 did not elevate extracellular glutamate levels in rats treated with EAAT3
antisense oligonucleotides. These animals did. however, experience epileptic seizures and
slightly impaired motor skills in some cases. These data support the findings that
EAAT1 and EAAT2 play a crucial role in synaptic clearance of glutamate, however, the
role of EAAT3 in preventing neurological damage seems to be unclear due to conflicting
data.
Evidence suggests that EAAT2 is responsible for the greatest amount of cerebral
glutamate transport (Robinson, 1991). When EAAT2 was knocked out by homologous
recombination, levels of residual glutamate increased in the brain and the mice suffered
lethal spontaneous seizures (Tanaka, 1997). Both the antisense and knockout studies are
consistent with the data documenting the loss of EAAT2 as being the major cause of the
motor neuron degeneration that plagues patients with ALS (Rothstein. 1995).
Conversely, when EAAT3 (EAAC1) knockout mice were produced, no
neurodegeneration was observed (Peghini and Stoffel, 1997). Instead, the mice
developed dicarboxylic aminoacidurea, analogous to an inborn error of glutamate and


Figure 4-5. Expression of the GFP(C3)-CAT1 fusion protein in normal and LPI human
fibroblasts. Normal (A) and LPI (B) fibroblasts were transfected for 3 h with the
GFP(C3)-CAT1 fusion protein according to the lipofectamine protocol described in the
Methods Chapter. Following 24-48 h of expression, cells were fixed with -20C MeOH
and visualized by deconvolution microscopy. Images were processed from three
independent transfections and the fluorescence patterns of expressed proteins were
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.


71
Lvsine Transport in Whole Cells and Plasma Membrane Vesicles of LPI
System y" is a Na-independent activity that catalyzes facilitated transport and
exchange reactions (White. 1985), but also may permit accumulation of cationic amino
acids against a concentration gradient driven by the trans-membrane potential (Bussolati
et al., 1987). Under normal physiological conditions. System y~ mediates the uptake of
cationic amino acids such as arginine, lysine, and ornithine. CAT1 was the first cDNA to
be identified that exhibited System y* activity (Albritton et al., 1989). This transporter,
which also serves as the murine ecotropic leukemia virus receptor, mediates the Na -
independent, high-affinity transport of cationic amino acids in most mammalian cells
(Kim et al., 1991; Wang et al., 1991). The second member of the "CAT" family to be
isolated, CAT2. shares 61% identity with CAT1 at the amino acid level, and mediates
Na -independent, high-affinity transport of cationic amino acids in activated T
lymphocytes (MacLeod et al., 1990). The third clone to be identified, CAT2a, is the
result of alternative splicing of the CAT2 gene. It differs from CAT2 by only 41 amino
acids, yet it exhibits a 10-fold lower affinity for arginine and is liver-specific (Closs et al.,
1993). The last member of the "CAT" family, CAT3, shares the greatest homology with
CAT1 and mediates the Na-independent high-affinity transport of cationic amino acids
in the brain (Hosokawa et al., 1997; Ito and Groudine, 1997).
In cultured fibroblasts, there is little difference between the initial rates (measured
after 15 sec) of lysine uptake in normal and LPI cells, however, after 20 min, the LPI
cells accumulate two-fold more H-lysine than control cells (Handlogten and Kilberg,


43
antibodies were purchased from Chemicon (EAAT1-C) and a-Diagnostics (EAAT1-D).
Both antibodies were incubated with MeOH-fixed human fibroblasts according to the
same procedures used for the EAAT1-R and EAAT1-S antibodies. Nuclear staining was
observed using a 1:1000 dilution of the EAAT1-C antibody (Figure 3-4 B), detected with
a 1:400 dilution of goat anti-guinea pig IgG linked to Cy3. but that labeling associated
with the nuclear membrane was fainter than with the EAAT1-R. and there also was
significant staining of a cellular vesicle population. Thus, the pattern of fluorescence
generated by the EAAT1-C antibody appeared to be a mixture of the results obtained
from the EAAT1-R and EAAT1-S antibodies. No specific or background staining was
detected when human fibroblasts were incubated with EAAT1-C priman' or anti-guinea
pig secondary antibodies alone. However, attempts to enhance the EAAT1-C nuclear-
specific staining by centrifuging (5 min at 13.000 x g) or filtering the EAAT1-C antibody
through a 2 pm pore syringe prior to use failed. A 1:50 dilution of the EAAT1 -D
antibody resulted in a uniform, diffuse cytosolic labeling with no specific staining of the
nuclear membrane, cytoplasmic vesicles, or other cellular structures (data not shown).
Localization of the EAAT1 glutamate transporter in other cell types. To
determine whether or not the nuclear localization of EAAT1 was unique to fibroblasts,
several other cell types were labeled with both the EAAT1-R and EAAT1-S antibodies.
Hela. HepG2 (human hepatoma), and pulmonary artery endothelial cells (PAEC) were
each fixed with -20C MeOH and stained with 1:50 dilutions of both EAAT1-R and
EAAT1-S according to the immunofluorescence assay described in the Methods Chapter.
The EAAT1-R antibody labeled the nuclei of both Hela (Figure 3-5 A) and HepG2 (data


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SUBCELLULAR LOCALIZATION AND TRAFFICKING
OF AMINO ACID TRANSPORTERS
By
Kelly Kristin McDonald
August, 1998
Chairman: Dr. Michael S. Kilberg
Major Department: Biochemistry and Molecular Biology
Mammalian amino acid transporters have been well-characterized with regard to
substrate specificity, kinetic parameters, and metabolic regulation. However, little
information is known concerning the cell biology or "life cycle" of plasma membrane
amino acid transporters. In this study, molecular and cell biological techniques, including
immunohistochemistry, transporter mutants and expression of green fluorescent protein
fusions, and deconvolution microscopy, were used to examine the cellular localization
and trafficking of specific amino acid transporters under normal and diseased conditions.
The availability of cDNAs and antibodies for specific members of the EAAT anionic and
the CAT cationic amino acid transporter families have provided an avenue for comparing
the subcellular distribution of amino acid transporters from the same, as well as different
gene families. Although similar in structure and function, the EAAT1 glutamate
transporter was detected primarily in the nuclear membrane in certain cell types, such as
IX


155


Figure 4-8. Intracellular staining of normal and LPI fibroblasts with antibodies against
proteins of the endoplasmic reticulum and Golgi Complex. According to the protocol
described in the Methods Chapter, normal (A, C, and E) and LPI (B, D, and F) fibroblasts
were fixed with -20C MeOH and subjected to immunohistochemistry with the anti-
KDEL antibody specific for the endoplasmic reticulum (A and B), the anti-
sialyltransferase (ST) antibody specific for the Golgi and Trans-Golgi Network (TGN) (C
and D), and the anti-mannose 6-phosphate receptor (M6PR) antibody specific for the
TGN and late endosomes (E and F). The anti-KDEL primary antibody was visualized
with a goat anti-mouse IgG conjugated to FITC, and the anti-ST and anti-M6PR
antibodies were detected with a goat anti-rabbit IgG linked to FITC. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. The data shown represent analysis of 0.2 pm sections through the cells.


92
acidic in nature. The above observations all suggest that abnormal lysosomal-like
structures may be one of the problems contributing to the LPI disorder.
Interestingly, the LPI fibroblasts exhibit several morphological characteristics that
appear in the fibroblasts of patients with Chediak-Higashi Syndrome (CHS). CHS is an
autosomal recessive disorder that is characterized by clusters of giant lysosomes
concentrated in the perinuclear region of several cell types, including fibroblasts (Jones et
al., 1992). The gene responsible for this disorder has been cloned, yet little is known
about the function of the protein it encodes. The Beige protein (the mouse homologue to
the CHS gene) is believed to be a 400 kDa cytosolic protein that is expressed in most
mouse tissues (Perou et al., 1997). A deficiency in the Beige protein leads to the
formation of abnormally large lysosomes, whereas overexpression of the protein results
in unusually small lysosomes. This finding, along with results from a lysosome-
lysosome fusion assay (Ward et al., 1997) suggests that the CHS/Beige protein plays a
role in the fission (and not fusion) of lysosomal vesicles. However, there is no evidence
that the CHS/Beige protein interacts with lysosomes, or any other organellar membranes
(Perou et al., 1997). The CHS/Beige protein contains no obvious signal sequences,
transmembrane spanning domains, or membrane-anchors. The presence of WD40 repeats
(a known protein interaction domain) within the CHS/Beige protein sequence suggests
that it may interact with other proteins. Therefore, the CHS/Beige protein may
participate in lysosomal fission indirectly, via the interaction of another protein.
No CHS/Beige-associated proteins have been identified, however, preliminary
experiments by Davies et al. demonstrate that the down-regulation of either Rab7 or Rab9
proteins induces fibroblasts to form large vacuoles resembling those of the CHS


50
sequence, did not stain the nucleus would be consistent with the hypothesis that the
EAAT1-R antibody recognized a different, yet related, protein to EAAT1. Also, if a
truncated form of EAAT1 resides in the nucleus, then it is possible that it would be
recognized by the EAAT1-R but not the EAAT1-S antibody. This later explanation,
however, is not supported by the data from the immunoblot analysis in which the
EAAT1-R antibody recognized a nuclear protein corresponding to the molecular weight
of the intact EAAT1 protein.
The third explanation is that the EAAT1-R antibody is detecting a protein that is
completely different from the EAAT1 glutamate transporter, but shares a few amino acids
in common. The peptide inhibition indicates that the EAAT1-R antibody interaction is
specific, therefore, the unrelated nuclear protein would most likely have at least one
homologous epitope to EAAT1. BLAST database searches did not identify any proteins
other than EAAT/ASCT family members, with significant homology to EAAT1, but such
searches do not always reveal small stretches of common sequence. Also, it would be
quite coincidental that the unknown protein has the same molecular mass as EAAT1 on
an SDS-PAGE gel. The fact that the EAAT1-R antibody showed very little co
localization with the EAAT1-GFP(N3) fusion protein is puzzling. It is surprising that the
EAAT1-R antibody would recognize an unrelated or homologous protein, and not detect
the expressed EAAT1-GFP(N3). However, the fact that EAAT1-S recognized the
EAAT1-GFP(N3), suggests that the EAAT1-S antibody is specific for EAAT1, and does
not recognize those transporters or related proteins localized to the nucleus.
Although preliminary immunoblot analysis supports the immunofluorescent
nuclear localization of EAAT1, more extensive studies need to be performed. Various


Figure 4-7. Intracellular staining of normal and LPI fibroblasts with antibodies against
plasma membrane and cytoskeletal proteins. According to the protocol described in the
Methods Chapter, normal (A, C, and E) and LPI (B, D, and F) fibroblasts were fixed with
-20C MeOH and subjected to immunohistochemistry with the anti-P-integrin antibody
specific for the plasma membrane (A and B), the anti-caveolin antibody specific for
caveolae (C and D), and the anti-p-tubulin antibody specific for microtubules (E and F).
The anti-P-integrin and anti-P-tubulin primary antibodies were visualized with a goat
anti-mouse IgG conjugated to FITC, and the anti-caveolin antibody was detected with a
goat anti-rabbit IgG linked to FITC. Staining from three independent experiments was
analyzed by deconvolution microscopy and shown to be reproducible. The data shown
represent analysis of 0.2 pm sections through the cells.


77
LPI-derived vacuoles have been detected by electron microscopy in the cell lines from
two different LPI patients (Woodard and Kilberg. unpublished data; McDonald and
Kilberg, unpublished data). The only previous microscopic observation reported was in a
paper by Simell and coworkers who mentioned apparent alterations in the architecture of
hepatocytes (Simell et al., 1975). They reported an accumulation of vesicles containing a
"fibrillogranular material." although it is not known whether these represent the same
vesicles we have identified.
CAT1 antibody production against human. The preliminary immunofluorescence
research was performed using a CAT1 polyclonal antibody that was generated against a
25 amino acid sequence (SIKNWQLTEKNFSCNNNDTNVKYGE) from the third
extracellular loop of the murine CAT1 sequence (Woodard et al., 1994). The murine
CAT1 antibody was shown to stain specifically a wide variety of cell types from several
species and was inhibited by pre-incubation with the corresponding peptide. However,
this anti-murine CAT1 antibody did not immunoblot well, and it is possible that it is not
optimal for detection of denatured protein in the human cell lines being tested. For this
reason, a polyclonal antibody was generated against a sequence of 20 amino acids
(CEEASLDADQARTPDGNLDQ) at an intracellular site of the human CAT1 protein
(Cocalico Biologicals, Inc.. Reamstown, PA). Enzyme-linked immunosorbent assays
(ELISA) were performed on the individual serum samples to determine antibody titers.
Briefly. 96-well plates were coated with the human CAT1 peptide (above), and incubated
with a 1:50 to a 1:32,000 dilution of the immune or pre-immune serum collected from the
inoculated rabbit. Following the removal of the primary antibody, an alkaline-
phosphatase conjugated secondary antibody was added to the plates and detected with a


151


CHAPTER 2
MATERIALS AND METHODS
This Chapter contains the general materials and methods used during the course of
this research project. Chapter 3 and Chapter 5 include additional Methods Sections that
are specific for the work described in those chapters.
Materials
Fibronectin, bovine serum albumin (BSA), Triton X-100, and Triton X-l 14,
kanamycin, nocodozole, dimethyl sulfoxide, and adenosine triphosphate (ATP) were
purchased from Sigma Chemical Company (St. Louis, MO). RPMI and MEM media,
goat serum (NGS). fetal bovine serum (FBS), lipofectamine, OptiMEM, EcoRI and
Hindlll restriction enzymes and buffers, T4 polynucleotide kinase and buffer, and all
PCR primers and mutagenesis oligonucleotides were purchased from Gibco BRL
(Gaithersburg. MD). Paraformaldehyde (PFA), glycine and methanol (MeOH) were
purchased from Fisher Scientific (Pittsburgh, PA). Fluoromount-G was purchased from
Southern Biotechnology Associates (Birmingham, AL). The MORPH mutagenesis kit
was obtained from 5 Prime 3 Prime (Boulder, CO), the pCR2.1 TA cloning kit was
obtained from InVitrogen (Carlsbad, CA), and the DNA gel purification kit and DNA
plasmid purification kits were obtained from Quiagen (Valencia, CA.). The
nitrocellulose membranes were purchased from Cuno, Inc. (Meridian, CT) and the
enhanced chemiluminesence reagents were purchased from Pierce (Rockford, IL). The
17


35
In a study of the activity and localization of the glutamate transporters in day 14
and day 20 rat placenta. Matthews et al. detected EAAT1 in the nuclei of the maternal
decidua and placental trophoblast (Matthews et al.. 1998). In other cell types of the
placenta, however. EAAT1 was detected on the plasma membrane and in one or more
intracellular vesicle populations. EAAT1. EAAT2, and EAAT3 mRNA and protein
expression were increased in the day 20 placenta, and the expression patterns were cell-
type specific for each isoform. Neither EAAT2 nor EAAT3 was observed in the nuclei of
any of the cell types examined by immunohistochemistry. Although EAAT4 mRNA was
identified in both day 14 and day 20 placenta. EAAT4 protein was not detectable in either
sample by immunoblot or immunohistochemistry.
Over the past two decades, extensive research has been performed to determine
the ionic requirements, substrate specificity and kinetic parameters of the glutamate
transporters, however, there is almost no information regarding intracellular pools,
trafficking through compartments, or protein arrangement on the plasma membrane.
One of my intentions for this project was to answer some basic questions regarding the
cellular and subcellular distribution of the different isoforms. Using sequence-specific
antibodies generated against the individual EAAT transporters. I could explore the
expression of the isoforms in cell lines other than the brain. In addition, I could
determine if the different EAAT family members display a similar or distinct staining
pattern in regard to abundance and distribution of protein. This chapter describes the
results obtained from staining human fibroblasts with antibodies against the glutamate
transporters EAAT1 and EAAT3. More extensive work involving co-localization
experiments in several cell lines was performed using EAAT1 antibodies from several


175
Chen, H., Hughes, D. D., Chan, T., Sedat, J. W., and Agard, D. A. (1996) J. Structural
Biol. 116,56-60
Chen, H., Sedat, J. W., and Agard, D. A. (1989) in Handbook of Biological Confocal
Microscopy pp. 127-135,1.M.R. Press, Madison, WI
Cheng, S. H., Gregory, R. J., Marshall, J., Paul, S., Souza, D. W., White, G. A.,
O'Riordan, C. R. and Smith, A. E. (1990) Cell 63, 827-834
Chiles, T. C., O'Brien, T. W. and Kilberg M. S. (1987) Anal. Biochem. 163, 136-142
Cho, K. 0., Hunt, C. A. and Kennedy, M. B. (1992) Neuron 9, 929-942
Chou, H.-F., Vadgama, J. and Jonas, A. J. (1992) Chem. Med. Metabol. Biol. 48, 179-
193
Chun, M., Liyanage, U. K., Lisanti, M. P., and Lodish, H. F. (1994) Proc. Natl. Acad.
Sci, U. S. A. 91, 11728-11732
Clements, J. D Lester, R. A. J., Tong, G., Jahr, C. E. and Westbrook, (1992) Science
258, 1498-1501
Closs, E. I. (1996) Amino Acids 11, 193-208
Closs, E. I., Albritton, L. M. Kim, J. W. and Cunningham, J. M. (1993) J. Biol. Chem.
268, 7538-7544
Colley, N. J Baker, E. K Stamnes, M. A. and Zuker, C. S. (1991) Cell 67, 255-263
Conrad, P. A., Smart, E. J., Ying, Y-S., Anderson, R. G. W., and Bloom, G. (1995) J.
Cell Biol. 131,1421-1433
Conradt, M. and Stoffel, W. (1997) J. ofNeurochem. 68, 1244-1251
Conradt, M Storck, T and Stoffel, W. (1995) Eur. J. Biochem. 229, 682-687
Cooke, J. P Andn, N. A., Girerd, X. J., Hirsch, A. T., and Creager, M. A. (1991)
Circulation 83, 1057-1062
Cormack, B. P Valdivia, R. and Falkow, S. (1996) Gene 173, 33-38
Cowbum, R., Hardy, J., Roberts, P. and Briggs, R. (1988) Neurosci. Lett. 86, 109-113
Creager, M. A., Gallagher, S. J., Girerd, X. J., Coleman, S. M., Dzau, V. J., and Cooke, J.
P. (1992) J. Clin. Invest. 90, 1248-1253
Dall'Asta, V., Gazzola, G. C., Franchi-Gazzola, R., Bussolati, O., Longo, N. and Guidotti,
G. G. (1983) J. Biol. Chem. 258,6371-6379


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Michael S. Kilberg, Chair (J
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
i
Brian D. Cain
Associate Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor of Philosophy.
William A. Dunn
Associate Professor of Anatomy and
Cell Biology
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
"Susan C. Frost
Associate Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
phy/' \
Peter M. McGuire
Assistant Professor of Biochemistry
and Molecular Biology


159


12
individual steps that contribute to the "life cycle" of plasma membrane proteins have been
documented. Although little information is available, it is likely that many similarities
exist between the "life cycle" of amino acid transporters and other plasma membrane
proteins. From biogenesis to degradation, these proteins are transported through a
complex system of membrane compartments and organelles by specific vesicles that bud
from a donor membrane and fuse with a target membrane (Ivessa et al., 1995). A
combination of coat proteins and several classes of monomeric GTPases (i.e., Rabs) are
believed to regulate certain steps in vesicle trafficking. The SNARE hypothesis describes
the mechanism by which transport vesicles target membranes (Alberts et ah, 1994). v-
SNAREs are proteins on the vesicle membranes and t-SNAREs reside on the target
membranes. v-SNAREs and t-SNAREs are suspected to function as structural proteins
that interact at the point of vesicle docking. Over 30 different Rab proteins have been
identified and each is believed to play some role to ensure the specificity of individual
vesicle docking/fusion events of the membrane trafficking pathways (Nuoffer and Balch,
1994).
Three primary trafficking pathways have been described for various membrane
proteins: the biosynthetic-secretory pathway, the endocytic-exocytic pathway, and the
degradative pathway. In the biosynthetic pathway, integral membrane proteins and
secretory proteins are co-translationally inserted into the membrane or the lumen,
respectively, of the endoplasmic reticulum (ER) where early oligosaccharide modification
and proteolytic processing begin. Select proteins receive fatty acylation moieties either
during or following, the translation event (reviewed by Solski et al., 1995). Myristic and
palmitic acid modifications contribute to membrane binding and stability, and more


31
L-Glutamate and L-aspartate are important nutritional substances that contribute
to a variety of biochemical pathways in the brain and peripheral tissues. Specialized
glutamatergic neurons in the brain produce and store glutamate, the major excitatory
neurotransmitter in the mammalian central nervous system, until it is released into the
synaptic cleft in response to different stimuli. Intracellular concentrations of glutamate in
the brain reach approximately 10 mM, with the highest concentrations at nerve terminals
(Shupliakov et al., 1992; Storm-Mathisen et ah. 1992). Extracellular levels of glutamate
are carefully maintained below 3 pM, except during neurotransmission of a signal when
concentrations may reach between 1-2 mM (Nicholls. 1993; Clements et ah. 1992).
Members of the EAAT family of transporters, primarily the glial-specific EAAT1 and
EAAT2 (Rothstein et ah, 1996), are responsible for the high-affinity Na-dependent
transport of glutamate out of the synaptic cleft against a thousand-fold concentration
gradient. This carefully regulated transport activity is crucial for maintaining glutamate
concentrations below the level that is toxic to neurons.
Glutamatergic transmission is believed to contribute to normal brain activities
such as learning and memory (Bliss et ah, 1993), however, elevated levels of extracellular
glutamate are neurotoxic and can lead to several neuro-degenerative diseases such as
Amyotrophic Lateral Sclerosis (ALS), Huntingtons disease, and probably Alzheimers.
In 1992. Rothstein et ah showed that brain and spinal cord samples taken from the
autopsies of ALS patients revealed a reduced level of glutamate uptake (Rothstein et al.,
1992). More recently it was determined that the decrease in glutamate uptake in some
cases of sporadic ALS is due to the selective loss of the astroglial EAAT2 glutamate


Figure 3-7. Expression of GFP and GFP-EAAT1 fusion proteins in human fibroblasts.
Human fibroblasts were transfected for 3 h with GFP(N3) only (A), or with the EAAT1-
GFP(N3) (B), and GFP(C3)-EAAT1 (C) fusion proteins according to the lipofectamine
protocol described in the Methods Chapter. Following 24-48 h of expression, cells were
fixed with -20C MeOH and visualized by deconvolution microscopy. Images were
processed from three independent transfections and the fluorescence patterns of expressed
proteins were determined to be reproducible. The data shown represent analysis of 0.2
pm sections through the cells.


Figure 5-6. Immunofluorescent staining of PAEC transfected with GFP(C3)-CAT1
fusion protein. PAEC were transfected for 3 h with the GFP(C3)-CAT1 fusion protein
according to the lipofectamine protocol described in the Methods Chapter. Following 24-
48 h of expression, cells were fixed with -20C MeOH and stained with antibodies
against caveolin (A) and eNOS (B). Both the caveolin and eNOS antibodies were
detected with a goat anti-rabbit IgG conjugated to Texas Red and visualized by
deconvolution microscopy. Images were processed from three independent experiments
and the staining was determined to be reproducible. The data shown represent analysis of
1.0 pm sections through the cells. One cell in panel B is outlined in order to distinguish
the PAEC periphery.


166


141
with the addition of a 1 h incubation of PAEC membrane vesicles with 5 mM DMS prior
to the solubilization of plasma membrane proteins. Similar results, in which no co
precipitation of eNOS with CAT1, were obtained from immunoprecipitations performed
after treatment with the cross-linking reagent, DMS (data not shown).
Mutagenesis of potential caveolar targeting sequences in CAT1. The following
experiments were performed to investigate the mechanism by which CAT1 is localized to
caveolae. The Introduction to this chapter discusses a variety of proteins that are subject
to myristoylation, palmitoylation, or a combination of both as targeting signals for
caveolar localization. CAT1 has a glycine at position 2 and three cysteines at positions 3,
20, and 30 that are conserved between the mouse and human transporter, and, as
described previously, corresponding residues in other proteins, including eNOS, are
required for caveolar localization. To determine if these residues are necessary for
targeting CAT1 to caveolae, three mutants were constructed (see Methods Chapter), the
cDNAs expressed in PAEC, and cells assayed for co-localization with anti-caveolin and
anti-eNOS antibodies. The following mutations were made using the GFP(C3)-CAT1 as
the template (see Methods Chapter), and confirmed by restriction analysis and sequencing
(ICBR DNA Sequencing Core): CATMUT1, cysteines at 20 and 30 changed to serines;
CATMUT2, glycine at position 2 was changed to alanine; CATMUT3, glycine at
position 2 changed to alanine and cysteines at positions 3, 20, and 30 changed to serine.
Wild-type GFP(C3)-CAT1 and each of the mutants were transfected into PAEC as
described in the Methods Chapter, fixed with -20C MeOH after 24 to 48 h of expression,
and stained with anti-caveolin or anti-eNOS antibodies. The caveolin and eNOS
antibodies were detected using a goat anti-mouse IgG conjugated to Texas Red. As


181
Pollock, J. S., Forstermann, U., Mitchell, J. A., Warner, T. D., Schmidt, H. H., Nakane,
M and Murad, F. (1991) Proc. Natl. Acad. Sci. USA 88, 10480-10484
Rajantie, J., Simell, O. and Perheentupa, J. (1981) J Clin. Invest. 67, 1078-1082
Rakowicz-Szulczynska, E. M., Linnenbach, A. J. and Koprowski, H. (1989) Mol.
Carcinog. 2, 47-58
Ravichandran, L. V. and Johns, R. A. (1995) FEES Lett. 374, 295-298
Robbins, E. and Marcus, P. I. (1963) J. Cell Biol. 18, 237-250
Robinson, M. B., Hunter-Ensor, M. and Sinor, J. (1991) Brain Res. 544, 196-201
Rossitch, E., Jr., Alexander, E., Black, P. M., and Cooke, J. P. (1991) J. Clin. Invest. 87,
1295-1299
Rothstein, J. D., Dykes-Hoberg, M., Pardo, C. A., Bristol, L. A., Jin, L., Kunci, R. W.,
Kanai, Y., Flediger, M. A., Wang, Y., Schielke, J. P. and Welty, D. F. (1996)
Neuron 16, 675-686
Rothstein, J. D., Martin, L. and Kunci, R. W. (1992) N. Engl. J. Med. 326, 1464-1468
Rothstein, J. D., Van Kammen, M., Levey, A. I., Martin, L. and Kunci, R. W. (1995)
Ann. Neurol. 38, 73-84
Schneider, J. A. and Schulman, J. D. (1982) in The Metabolic Basis of Inherited Disease
(Scriver, C. R., Beaudet, A. L., Sly, W. S. and Valle, D., ed) pp. 1844-1866,
McGraw-Hill, New York
Schnitzer, J. E Oh, P., Pinney, E and Allard, J. (1994) J. Cell Biol. 127, 1217-1232
Scott, H. L., Tannenberg, A. E. G. and Dodd, P. R. (1995) J. Neurochem. 64, 2193-2202
Segal, S. and Thier, S. O. (1989) in The Metabolic Basis of Inherited Disease (Scriver,
C. R., Beaudet, A. L., Sly, W. S. and Valle, D., ed) pp. 2479-2496, McGraw-Hill,
New York
Sessa, W. C., Garcia-Cardena, G., Liu, J., Keh, A., Pollock, J. S., Bradley, J., Thiru, S.,
Braverman, I. M. and Desai, K. M. (1995) J. Biol. Chem. 270, 17641-17644
Shafquat, S., Tamarappoo, B. K., Kilberg, M. S., Puranam, R. S., McNamara, J. O.,
Guadano-Ferraz, A. and Fremeau, R. T. (1993) J. Biol. Chem. 268, 15351-15355
Shaul, P. W., Smart, E. J., Robinson, L. J., German, Z., Yuhanna, I. S., Ying, Y.,
Anderson, R. G., and Michel, T. (1996) Biol. Chem. 271, 6518-6522


Discussion 143
CHAPTER 6 CONCLUSIONS AND FUTURE DIRECTIONS 169
LITERATURE CITED 173
BIOGRAPHICAL SKETCH 185
IV


LITERATURE CITED
Agard, D. A. (1984) Ann. Rev. Biophys. Bioeng. 13, 191-219
Agard, D. A., Hiraoka, Y Shaw, P. and Sedat, J. W. (1989) Methods Cell Biol. 30,
353-377
Agard, D. A. and Sedat, J. W. (1983) Nature 302, 676-681
Aisaka, K., Gross, S. S., Griffith, O. W., and Levi, R. (1989) Biochem. Biophys. Res.
Commun. 163,710-717
Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J. D. (1994) in
Molecular Biology of the Cell, third edition, Garland Publishing, Inc., New York
Albritton, L. M., Kim, J. W., Tseng, L. and Cunningham, J. M. (1993) J. Virol. 67,
2091-2096
Albritton, L. M., Tseng, L., Scadden, D. and Cunningham, J. M. (1989) Cell 57, 659-666
Amara, S. G. and Kuhar, M. J. (1993) Rev. Neurosci. 16, 73-93
Amato, A., Barbour, B., Szatkowski, M. and Attwell, D. (1992) J. Physiol. 479, 371-380
Amato, A., Barbour, B., Szatkowski, M. and Attwell, D. (1994) J. Physiol. Lond. 479,
371-380
Anderson, R. G. W. (1993) Trends Cell Biol. 3, 69-72
Aris, J. P. and Blobel, G. (1988) J. Cell Biol. 107, 17-31
Amal, J.-F., Yamin, J., Dockery, S., and Harrison, D. G. (1994) Am. J. Physiol. 267,
C1381 -C1388
Arriza, J. L., Eliasof, S., Kavanaugh, M. P. and Amara, S. G. (1997) Proc. Natl. Acad.
Sci. U.S.A. 94,4155-4160
Arriza, J. L., Fairman, W. A., Wadiche, J. L, Murdoch, G. H., Kavanaugh, M. P. and
Amara, S. G. (1994) J. Neurosci. 14,5559-5569
Arriza, J. L., Kavanaugh, M. P., Fairman, W. A., Wu, Y. N., Murdoch, G. H., North, R.
A. and Amara, S. G. (1993) J. Biol. Chem. 268, 15329-15332
173


113


133
using deconvolution microscopy as a means of obtaining high-resolution data analysis
(McDonald et al., 1997). The CAT1 transporter is not uniformly distributed over the
entire cell surface. Instead, the cell surface has discrete regions that contain a high
transporter content. Figure 5-1 (A and B) shows PAEC stained with a 1:50 dilution of
anti-CATl antibody detected with a 1:200 dilution of goat anti-rabbit IgG linked FITC.
Staining with the anti-CATl transporter antibody was completely inhibited by
preadsorption of the antibody with the corresponding peptide antigen (Fig. 5-1 C),
whereas incubation of the antibody with non-antigen peptide sequences from within the
CAT1 transporter did not block cell staining (data not shown).
Co-localization of CAT1 and caveolin within PAEC. To determine if the
observed clustering is a result of CAT1 localization to caveolae, PAEC were fixed with a
4% paraformaldehyde solution and stained with a 1:50 dilution of anti-CATl (Figure 5-2
A), a 1:50 dilution of anti-caveolin (Figure 5-2 B), or a combination of both antibodies
(Figure 5-2 C). The CAT1 antibody was detected using a 1:200 dilution of goat anti
rabbit IgG conjugated to FITC, and the caveolin antibody was visualized with a 1:200
dilution of goat anti-mouse IgG linked to Texas Red. Immunohistochemistry of the
PAEC with antibodies specific for caveolin documented intensely staining clusters,
consistent with proposed caveolae structure and localization. Co-staining of the PAEC
with anti-CATl and anti-caveolin antibodies resulted in significant overlap with regard to
the localization of the two proteins (Figure 5-2 C). It is clear from these results that the
majority of the clusters containing the arginine transporter coincide with caveolae in the
PAEC.


180
Nuoffer, C. and Balch, W. E. (1994)Annu. Rev. Biochem. 63, 949-990
O'Brien, H. A., Young, H. N., Pobey, J. M. and Furness, J. V. (1995) Histochemistry
103, 221-225
Okamoto, T., Schlegel, A., Scherer, P. E. and Lisanti, M. P. (1998) J. Biol. Chem. 273,
5419-5422
Otori, Y., Shimada, S., Tanaka, K., Ishimoto, I., Tao, Y. and Tohyama, M. (1994) Mol.
Brain Res. 27, 310-314
Oxender, D. L. and Christensen, H. N. (1963) J. Biol. Chem. 238, 3686-3699
Paddy, M. R., Belmont, A. S., Saumweber, H., Agard, D. A. and Sedat, J. W. (1990)
Cell 62,89-106
Palacin, M. (1994) J. Exp. Biol. 196, 123-137
Palmer, R. M. and Moneada, S. (1989) Biochem. Biophys. Res. Commun. 158, 348-352
Palmer, R. M., Rees, D. D., Ashton, D. S., and Moneada, S. (1988) Biochem. Biophys.
Res. Commun. 153, 1251-1256
Parton, R. G. (1996) Curr. Opin. Cell. Biol. 8, 542 -548
Pearce, I. A., Cambray-Deakin, M. A. and Burgoyne, R. D. (1987) FEBS Lett. 223, 143-
147
Peghini, P Janzen, J. and Stoffel, W. (1997) EMBO 16, 3822-3832
Pelham, G. R. (1991) Cell 67, 449-451
Perou, C. M., Leslie, J. D., Green W., Li, L., Ward, D. M. and Kaplan, J. (1997) J. Biol.
Chem. 272,29790-29794
Pind, S Riordan, J. R. and Williams, D. B. (1994) J. Biol. Chem. 269, 12784-12788
Pines, G., Danbolt, N. C., Bjoras, M., Zhang, Y, Bendahan, A., Eide, L., Koepsell, H.,
Storm-Mathisen, J., Seeberg, E. and Kanner, B. I. (1992) Nature 360, 464-467
Pines, G., Zhang, Y. and Kanner, B. I. (1995) J. Biol. Chem. 270, 17093-17097
Pisoni, R. L., Thoene, J. G. and Christensen, H. N. (1985) J. Biol. Chem. 260,4791-
4798
Podlecki, D. A., Smith, R. M., Kao, M., Tsai, P., Huecksteadt, T., Brandenburg, D.,
Lasher, R. S., Jarett, L. and Olefsky, J. M. (1987) J. Biol. Chem. 262, 3362-
3368


105


29
by 2- to 10-fold over that of the control uninjected oocytes (Arriza et al., 1997).
However, this was significantly less than the transport observed with EAAT1, EAAT2.
and EAAT3-expressing oocytes, which were reported to transport 50-fold over the level
of uninjected oocytes (Klockner et al.. 1993; Kanai et al.. 1995). Whereas
electrophysiological studies have shown that both EAAT4 and EAAT5 transport
glutamate poorly, they possess much stronger chloride channel properties only weakly
present in the other glutamate transporters (Fairman et al., 1995; Arriza et al., 1997).
Several laboratories have documented the involvement of glutamate in cell migration and
differentiation (Pearce et al., 1987; Mattson et al., 1988), as well as neuronal and
astroglial proliferation. Osnat Bar-Peled et al. have shown that each of the glutamate
transporters has a specific and unique distribution during brain development (Osnat Bar-
Peled et al., 1997), suggesting that the transporters play multiple functional roles during
brain maturation.
Three amino acid residues in the C-terminal sequence, the region of greatest
homology, have been identified as essential for glutamate transport activity (Conradt and
Stoffel, 1995; Pines et al., 1995). Using site-directed mutagenesis. Pines and coworkers
determined that aspartate 398, glutamate 404, and aspartate 470 are critical for EAAT2
activity, and glutamate 404 may contribute to substrate specificity (Pines et al., 1995).
Both aspartate 398 and 470 appear to be involved in transporter activity, rather than
stability or trafficking, and even the conservative replacement of glutamate abolishes
transport activity. Zhang et al. showed, using site-directed mutagenesis, that histidine
326 is required for glutamate transport by EAAT2, and probably contributes to the proton
translocation mechanism that accompanies the Na and K7-coupled transport activity


90
cationic amino acids across the basolateral plasma membranes of affected cells. This
transporter defect would provide an explanation for the increased intracellular
accumulation of cationic amino acids during LPI whole cell and vesicle transport assays.
Experiments were performed to test for the presence of CAT1, or other transporters, in
abnormal LPI vesicles using a higher-resolution microscope than had previously been
available. In addition, organelle-specific antibodies were used to compare the cellular
compartments involved in biosynthesis, endocytosis, and degradation in normal and LPI
cells. Immunofluorescence studies did not detect an association of CAT1 transporters
with the abnormal LPI-associated vesicle population previously reported. In addition, the
extracellular and intracellular distributions of the CAT1, EAAT1, EAAT3, and GLUT1
transporters, as well as the integrity of specific organelles involved in the trafficking
pathways of membrane proteins appeared normal in the LPI fibroblasts. Therefore, the
most important difference between the normal and LPI fibroblasts, detected by
immunofluorescence in this study, was an abnormal population of enlarged lysosomal-
like vacuoles (described above).
It is confusing as to why earlier immunofluorescence data documenting the
CAT 1-containing LPI vesicles could not be reproduced. However, several experimental
conditions were changed that may explain the apparent discrepancy. First, although the
primary cell line used in this study came from the same patient as before, the cells were
isolated and cultured at different times. Although it is unlikely, the presence of CAT1 in
the LPI vesicles could have involved a defect specific to the cell line used in the first set
of experiments. Second, the same sample of CAT1 antibody was not available for this
study. The CAT1 peptide and antibody were older during the experiments described


This dissertation was submitted to the Graduate Faculty of the College of
Medicine and to the Graduate School and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
August, 1998
Dean, College of Medicine
Dean, Graduate School


9
Amyotrophic Lateral Sclerosis (ALS), but the defective glutamate transport is not thought
to be the principle cause of the disease (Lin et al., 1998). The third member of the
growing glutamate transporter family, EAAC1, was cloned by oocyte expression cloning
using fractionated mRNA from rabbit intestine (Kanai and Hediger, 1992). Although
similar to GLAST1 and GLT1 in structure and transport properties, EAAC1 expression is
neuron-specific in the brain. EAAC1 transcripts have also been detected in small
intestine, kidney, liver, heart, placenta, and skeletal muscle (reviewed by Malandro and
Kilberg, 1996). In 1994, the human homolog to GLAST1 was identified (Arriza et al.,
1994; Kawakami et al., 1994) and given the name EAAT1, for Excitatory Amino Acid
Transporter 1. Shortly after, the human GLT1 sequence was reported (Arriza et al., 1994;
Manfras et al., 1994) and designated EAAT2, and the human homolog to EAAC1 was
cloned (Arriza et al., 1994; Kanai et al., 1995) and called EAAT3. It has since become
acceptable to refer to the glutamate transporters by the EAAT nomenclature regardless of
the species.
EAAT4 was isolated using degenerative oligonucleotide primers corresponding to
conserved sequences within the other members of the glutamate family (Fairman et al.,
1995). Northern analysis identified EAAT4 mRNA in the cerebellum and placenta, and
transport studies in oocytes demonstrated a high-affinity glutamate uptake that was
associated with chloride conductance. The final member of this family, EAAT5, was
cloned by Arriza et al. by screening a human retinal cDNA library with a glutamate
transporter cDNA isolated from salamander retina (Arriza et al., 1997). Like EAAT4,
EAAT5 may play a role in ion conductance instead of, or in addition to, providing
neurotransmitter clearance at the synaptic cleft. Electrophysiological studies of both


73
has been implicated as the defective cystine transporter in the inherited disease, cystinuria
(Calonge et ah. 1994).
System B0', first described by Van Winkle and colleagues (Van Winkle et ah.
1985), is responsible for cationic and neutral amino acid transport in various cell types,
including human fibroblasts. However, this is a Na*-dependent system, and therefore,
does not appear to be responsible for the increased LPI amino acid accumulation. System
y"L mediates the Na-independent high-affinity transport of cationic amino acids, as well
as the Na~-dependent high-affinity transport of L-leucine (Deves et ah, 1992). The
observation by our laboratory that 5 mM 2-amino-[2,2,l]-bicycloheptane-2-carboxylate
(BCH) inhibited leucine uptake in the normal and LPI vesicles is consistent with the
hypothesis that System y+L may be responsible for the unusual accumulation of cationic
and neutral amino acids in the LPI vesicles. L-leucine transport was inhibited by leucine,
lysine, arginine, and BCH in normal fibroblasts vesicles. However, lysine did not block
the L-leucine uptake in LPI vesicles. Collectively, the vesicle transport data is not
entirely consistent with the properties of any of the known transport systems. Therefore,
it is possible that an uncharacterized transport system is responsible for the activity
observed in the LPI vesicles. Alternatively, a known transport system that is altered in its
substrate specificity in the LPI cells may account for the elevated amino acid
accumulation.
Immunofluorescence Studies
When transport studies in LPI cells began, CAT1 was the only family member
that had been reported to mediate System y+ activity. Therefore, based on clinical


26
cleaved into small linear strands and will not be efficiently introduced and propagated
during the bacterial transformation. Following the digestion, the entire mutagenesis
reaction was added to 200 pi of E. coli MORPH mutS cells and incubated on ice for 20
min. The cells were heat-shocked at 42C for 2 min and spread on LB plates containing
30 pg/ml kanamycin for selection of the appropriate target plasmid. The mutS strain of
E. coli is used because it is deficient in DNA repair strand selection, therefore, it
randomly repairs either the mutant strand or the original template pair. This way, there is
a 50% chance that the mutant strand will be selected as correct and that sequence will be
propagated further. After selecting colonies and growing the bacteria in LB plus 30
pg/ml kanamycin, the plasmid DNA was isolated from several colonies with Quiagen
minipreps and digested using the enzyme specific to the engineered sequences to confirm
they contained the mutant strand (Sal I for CATMUT1 and Ava I for CATMUT2 and
CATMUT3). Large scale plasmid prep kits (manufactured by Quiagen) were used to
prepare the final mutant cDNAs. Each of the mutants was subjected to a series of
restriction digests to confirm the amino acid substitution(s) and to check for appropriate
sizes of the vectors and inserts.
Data analysis. Much of the data generated in the proposed experiments were
qualitative and required visual analysis. Each immunofluorescence experiment was
performed a minimum of three times to check for consistency among cell populations
plated on different days. In addition, normal and LPI cells of the same passage number
were assayed in parallel using the same reagents and antibody solutions. LPI cells from
several different patients were used for this study. Cells from the same patient were used


168


7
believed to function in transport, because most of the cloned transporters are thought to
span the membrane 12-14 times. That NBAT and 4F2hc have four or less trans
membrane domains has led to the hypothesis that they may serve as modulators, or
comprise only one subunit of a multimeric transporter complex (for review see Palacin,
1994). Co-precipitation and cross-linking experiments have recently provided strong
evidence that NBAT, a 90 kDa protein, is associated with a 50 kDa protein (Wang and
Tate, 1995). The NBAT protein is expressed in the microvilli of proximal tubules of the
kidney and the mucosa of the small intestine (Kanai et al., 1992). Expression of NBAT
in Xenopus oocytes has been shown to induce the high-affinity uptake of cystine. This
finding provided key evidence for NBAT's involvement in cystinuria, an autosomal
recessive disorder characterized by the hyperexcretion of cystine and cationic amino acids
into the urine (Segal and Thier, 1989; reviewed by Palacin, 1994). At least six distinct
missense mutations in NBAT have been documented in different patients with the
disease, but all result in defective transport of cystine through the epithelial cells of the
renal tubule and intestinal tract (Calonge et al., 1994).
Anionic Amino Acid Transport Systems
The anionic transporter family currently includes at least five members that
mediate glutamate/aspartate transport. These Na+-dependent transporters share a similar
structure with six predicted trans-membrane spanning domains in the N-terminal portions
and a large hydrophobic region at the C-termini that may represent additional trans
membrane domains. Each transporter has at least two putative glycosylation sites and
shares from 40-68% amino acid sequence identity with the other members of the family.


Figure 5-2. Co-localization of CAT1 and caveolin on PAEC. According to the
immunofluorescence protocol described in the Methods Chapter, PAEC were fixed with
4% PFA and stained with antibodies against CAT1 (A), caveolin (B), or both (C). Co
localization was assayed by simultaneously incubating PAEC with a polyclonal antibody
against CAT1 detected with goat anti-rabbit IgG linked to FITC and a monoclonal
antibody against caveolin detected with goat anti-mouse IgG conjugated to Texas Red.
Staining from three independent experiments was analyzed by deconvolution microscopy
and shown to be reproducible. The data shown represent analysis of 1.0 pm sections
through the cells.


137
(Figure 5-7 D), as described above. Cells were incubated with MEM + DMSO (0.83%),
as illustrated in Figure 5-7 C, to ensure that the DMSO was not responsible for effects
observed with nocodazole treatment. PAEC were also stained with a 1:100 dilution of
mouse anti-rat P-tubulin antibody following either DMSO (Figure 5-7 A) or nocodazole
(Figure 5-7 B) treatment to test for microtubule disruption. The tubulin was completely
disrupted following the 1 hr nocodazole treatment. This was determined based on the
shift of the fluorescent pattern from long, thin strands of microtubules to small punctate
free tubulin subunits. Five to ten images, collected from three individual experiments
indicated that the eNOS Golgi staining and the CAT1 "patches" were altered in the
nocodazole-treated cells. eNOS staining in the perinuclear region was disrupted, and the
eNOS-labeled vesicles were more diffusely distributed throughout the cytoplasm. The
CAT1 staining disappeared completely from the perinuclear region, and the "patches"
appeared to be smaller and almost entirely located at the cell periphery following
nocodazole treatment. The co-localization of CAT1 and eNOS in DMSO appeared to be
predominantly intracellular, and was completely disrupted following nocodazole
treatment. A sample of the PAEC were allowed to recover in MEM + 10% FBS for 3 h
after removing the nocodazole. These cells were then fixed with 4% paraformaldehyde
and stained with antibodies against CAT1 and eNOS. Three hours is long enough for the
re-polymerization of microtubules, and this was confirmed by staining nocodazole-treated
PAEC with anti-(3-tubulin antibody after the 3 h recovery period (data not shown). The
reformation of the CAT1 clusters, as well as co-localization of CAT1 and eNOS,
appeared to be complete after a 3 h recovery in MEM + 10% FBS (data not shown).


64


107


184
Woodard, M. H., Dunn, W. A., Laine, R. O., Malandro, M., McMahon, R., Simell, O.,
Block, E. R., and Kilberg, M. S., (1994) Am. J. Physiol. 266, E817-E824
Wu, J. Y., Robinson, D., Kung, H., and Hatzoglou, M. (1994) J. Virol., 68, 1615-1623
Yoshimoto, T., Yoshimoto, E. and Meruelo, D. (1991) Virology 185, 1-17
Zembowicz, A., Tang, J. and Wu, K. K. (1995) J. Biol. Chem. 270, 17006-17010
Zerangue, N. and Kavanaugh, M. P. (1996) Nature 383, 634-637
Zhang, Y., Pines, G. and Kanner, B. I. (1994) J. Biol. Chem. 269, 19573-19577
Zharikov, S. I. and Block, E. R. (1997) Biochim. Biophys. Acta, in revision


172
explanation for the "arginine paradox," an observation that the cellular eNOS
preferentially uses extracellular arginine for NO production despite the high, even
saturating, intracellular levels of the substrate. A physical interaction between CAT1 and
eNOS would also introduce a novel role for the CAT1 transporter in signal transduction
via the production of NO.
Several future experiments will be conducted to determine if an actual interaction
exists between CAT1 and eNOS. Initial co-precipitation experiments using the CAT1
antibody to precipitate eNOS failed. Therefore, future experiments will attempt to co
precipitate eNOS using an antibody against the GFP(C3)-CAT1 fusion protein. Using a
commercially available antibody generated against the GFP tag will provide an
alternative in case the CAT1 antibody is not suitable for the immunoprecipitation
protocol. Additional future experiments will include the examination of the intracellular
co-localization of CATl/10e6 and eNOS/sialyltransferase dual-staining experiments,
however, it is unclear whether an interaction exists in this compartment. Also, the
integrity of the CATl/eNOS complex following the disruption of actin and other
cytoskeletal components will be explored further.
The data presented in this thesis begin to answer some questions regarding the cell
biology of amino acid transporters. As the molecular tools improve, and as more
transporters are identified and used to generate antibodies, a more thorough investigation
of the "life cycle" of amino acid transporters will be possible.


101
10.0 um


41
double-labeled with a 1:200 dilution of the EAAT3 antibody and a 1:5 dilution of the
transferrin receptor (Figure 3-2 C). indicating that ven little of the EAAT3 glutamate
transporter is involved in recycling under standard culture conditions.
Localization of endogenous EAAT1 glutamate transporter in human fibroblasts bv
immunofluorescence. Initial experiments were performed using the EAAT1-R antibody
generated in Dr. Jeffrey Rothstein's laboratory against an N-terminal peptide sequence.
This antibody will be referred to as EAAT1-R (see Table 3-1 for details on EAAT1
antibodies). Unlike the "patches" observed on human fibroblasts with the anti-EAAT3
antibody, the anti-EAATl-R antibody did not detect any protein on the surface of
paraformaldehyde-fixed fibroblasts even at a 1:25 antibody dilution. When fibroblasts
were fixed with -20C MeOH and stained for intracellular EAAT1, using a 1:50 dilution
of the anti-EAATl-R antibody, the nuclear membrane and nuclear matrix were the
primary structures detected (Figure 3-3 A). Small vesicles throughout the cytoplasm,
resembling the vesicles labeled with the EAAT3 antibody, were also apparent. Staining
of the anti-EAATl-R transporter antibody was completely inhibited by preadsorption of
the antibody with 50 pg/ml of the corresponding peptide antigen for 12 h at 4C (Figure
3-3 B). To confirm that the EAAT1-R antibody was staining the nuclear membrane,
MeOH-fixed fibroblasts were double-labeled with antibodies against EAAT1-R (1:50
dilution) and 414(1:10 dilution) (Figure 3-3 C). The latter is an antibody generated
against an epitope shared by several proteins of the nuclear pore complex (Davis and
Blobel, 1986). There was significant co-localization of the EAAT1-R and 414
antibodies, strongly suggesting that EAAT1-R was labeling the nuclear membrane. On


Figure 4-12. Visualization of acidic compartments of normal and LPI fibroblasts with
acridine orange. Normal (A) and LPI (B) fibroblasts were treated with acridine orange
(AO) according to the protocol described in Chapter 4. Briefly, cells were incubated with
5 pg/ml of AO for 15 min before fixing with -20C MeOH and viewing with a Nikon
Axiophot epifluorescence inverted microscope. The digitized images were captured
using a Spot CCD camera with a resolution of 1315 x 1033 pixels (Diagnostics
Instruments, Inc., Sterling Heights, MI). Fluorescence from three independent
experiments was analyzed and shown to be reproducible.


LIST OF TABLES
Table page
1-1. cDNA Clones of Amino Acid Transporters in CAT and EAAT Families 2
3-1. Antibodies against Glutamate Transporters 39
4-1. Antibodies for Immunofluorescence Studies 78
5-1. Antibodies for Immunofluorescence Studies 130
5-2. Immunodepletion of CAT 1-mediated Arginine Transport Activity
by anti-eNOS Antibody 139
v


20
cells were incubated in a solution of PBS containing 20% normal goat serum (NGS) and
3% bovine serum albumin (BSA), for 1-2 h in order to block non-specific antibody
binding. For pre-immune or immune labeling, cover slips were removed from wells and
inverted onto 50 pi drops of pre-immune or primary antibody solution. The primary
antibodies were prepared in 20% NGS/PBS with 3% BSA (plus Triton-X when
appropriate), and for peptide competition assays, allowed to incubate with 50 pg/ml of
corresponding peptide overnight at 4C. All incubations were at room temperature
(unless otherwise indicated), and antibody reactions were performed on parafilm in a dark
humid box. The dark humid box was prepared by placing PBS-saturated gauze across the
bottom of a small Tupperware container. Following a 2 h pre-immune or primary
antibody incubation, cells were returned to the wells and washed three times with PBS.
The secondary antibody, prepared in 20% NGS/PBS with 3% BSA (plus 0.1-0.2%
Triton-X when appropriate), was applied in a similar manner to the primary, and allowed
to incubate for 1 h before unbound molecules were removed with three 5 min PBS
washes. Following the last wash, cover slips were mounted onto glass slides with a drop
of Fluoromount-G, allowed to dry, and the edges of the cover slip sealed with fingernail
polish.
The secondary antibodies used in the immunofluorescence assays were either goat
anti-mouse or goat anti-rabbit IgG conjugated to either Texas Red or FITC (fluorescein
isothiocyanate) fluorochromes (unless otherwise stated). Texas Red is excited when
exposed to a wavelength of 593 nm and emits a red fluorescence at a wavelength of 612
nm. FITC is excited at 494 nm and emits a green fluorescence at 517 nm. Because of the


91
above and may not have been as effective. As a result, new peptide should be prepared
and affinity-purified antibody generated for future experiments. Third, previous
immunofluorescence experiments were visualized using an epifluorescence microscope,
whereas the image processing for this project involved the use of a deconvolution
microscopy system. The ability to scan the cell in the z-plane, and the reduced
background provided by the deconvolution microscope, results in a higher degree of
resolution for antibody labeling. Dissection of the LPI cells in the z-plane may be
particularly important because of their tendency to shed pieces of the plasma membrane
and collect debris on their cell surface. Our laboratory has shown that this debris contains
CAT1 transport activity, stains with CAT1 antibody and thus, may create a problem for
detection of "intracellular" CAT1 when a standard epifluorescence microscope is being
used. The deconvolution microscope allows the user to non-invasively slice the cell into
sections from top to bottom. One or more sections can be taken from the middle of the
cell, thus avoiding fluorescent interference by cell surface debris.
Electron microscopy confirmed the presence of the large vacuoles previously
observed with light microscopy in the cytoplasm of LPI cells. Immunofluorescence
results indicate that the large vacuoles are probably lysosomes, or a related compartment
that contains cathepsin D and the lpgl 20 lysosomal membrane protein. When cells were
incubated with chloroquine, a reagent that neutralizes acidic compartments, the disease
phenotype of enlarged cytoplasmic vacuoles was induced in normal fibroblasts, and
exaggerated in LPI cells. Although lpgl 20 antibody staining of acridine orange loaded
cells was not possible, due to the rapid diffusion of the lysosomotropic drug, the pattern
of AO fluorescence indicated that the abnormal LPI lysosomes/vacuoles are probably


94
of the characteristics with the plasma membrane System y+. System c mediates the Na-
independent transport of cationic amino acids and exhibits the property of trans
stimulation. Cystinosis is an autosomal recessive disease that results from defective
efflux of cystine from the lysosomes (Schneider and Schulman, 1982). Administering the
therapeutic agent, cysteamine, is the primary mode of treatment for this disorder. This
therapy is presumed to be successful because System c is capable of lowering lysosomal
cystine concentrations by allowing efflux of the mixed disulfides, cysteine and
cysteamine, out of the lysosomes (Pisoni et al., 1985). It is possible that lysine or other
positively charged amino acids are accumulating in the lysosomes of LPI cells due to a
defect in the lysosomal System c cationic amino acid transporter. Abnormal
accumulation of amino acids could be responsible for the swelling of the lysosomes in the
LPI fibroblasts. It has been reported that the incubation of lysosomes with chloroquine
leads to a reduction in lysine efflux. This decreased amino acid efflux may explain why
incubating the normal and LPI fibroblasts with chloroquine resulted in the swelling of the
lysosomal-like vesicles. If the LPI lysosomes have an elevated amount of lysine due to a
partially defective cationic amino acid transporter, then blocking the efflux further, with
chloroquine, may lead to an exaggerated phenotype of the disease. In addition, blocking
the functional lysosomal transporter in the normal fibroblasts with chloroquine treatment
would induce the disease phenotype.
One of the future directions of this project may be to prove whether or not LPI is a
lysosomal disease. Lysosomes isolated from normal and LPI cells, by differential
centrifugation, can be measured by flow cytometry (Perou et al., 1997). In addition, the
lysosomal-enriched preparation can be used for measuring the influx and efflux of


95
cationic and neutral amino acids. This may indicate whether or not specific amino acids
are accumulating in, and leading to the enlargement of, lysosomes in the LPI cells.


45
steps required to obtain the 100.000 x g pellet, the greater the shift from monomer to
dimer/trimer forms of the transporter. Also, detection of the higher molecular mass
species of EAAT1 is consistent with published data from other laboratories documenting
the formation of glutamate transporter homomultimers (Haugeto et al., 1996).
When HepG2 cells were subjected to cell fractionation, SDS-PAGE, and
immunoblotting (see Methods section above), a 1:200 dilution of EAAT1-R was detected
in the 300 x g fraction in both the monomeric (approximately 70 kDa) and higher
molecular mass form (approximately 180 kDa) (Pappas and Kilberg, unpublished data).
EAAT1-R immunoreactivity was detected in the 180 kDa form in the 15.000 x g and
100,000 x g fractions, as well as two smaller species (approximately 70 and 75 kDa) in
the 100.000 x g fractions (data not shown).
Cell fractionation. SDS-PAGE. and immunoblotting were also performed using
PAEC according to the protocols in the Methods Section of this chapter. To determine if
the EAAT1 immunoreactivity in the nuclear (300 x g) fraction was a result of
contamination by other membranes or unbroken cells, a 1:100 dilution of A-l A5 ((3-
integrin) antibody was incubated with each of the fractions (data not shown). The A-1A5
antibody detects one or more p-integrin species (depending on the cell type) that migrate
at 210, 165, and 130 kDa (Hemler et al., 1984), and are specific for the plasma
membrane. I have confirmed the plasma membrane specificity of the antibody by SDS-
PAGE and immunoblot analysis (data not shown). Bands of approximately 165 kDa
were detected by the P-integrin antibody in the 300 x g fraction as well as the 15,000 x g
fraction of PAEC. Very little protein was observed in the 100,000 x g fraction. Although


Figure 3-8. EAAT1 immunofluorescent staining of human fibroblasts transfected with
EAAT1 -GFP(N3). Human fibroblasts were transfected for 3 h with the EAAT1 -
GFP(N3) fusion protein according to the lipofectamine protocol described in the Methods
Chapter. Following 24-48 h of expression, cells were fixed with -20C MeOH and
stained with antibodies against EAAT1-R (A) and EAAT1-S (B). Both EAAT1
antibodies were detected with a goat anti-rabbit IgG conjugated to Texas Red and
visualized by deconvolution microscopy. Images were processed from three independent
experiments and the staining was determined to be reproducible. The data shown
represent analysis of 0.2 pm sections through the cells.


Figure 5-3. Co-localization of CAT1 and eNOS on PAEC. According to the
immunofluorescence protocol described in the Methods Chapter, PAEC were fixed with
4% PFA and stained with antibodies against CAT1 (A), eNOS (B), or both (C). Co
localization was assayed by simultaneously incubating PAEC with a polyclonal antibody
against CAT1 detected with goat anti-rabbit IgG linked to FITC and a monoclonal
antibody against eNOS detected with goat anti-mouse IgG conjugated to Texas Red.
Staining from three independent experiments was analyzed by deconvolution microscopy
and shown to be reproducible. The data shown represent analysis of 1.0 pm sections
through the cells.


CHAPTER 1
INTRODUCTION
Overview of Mammalian Amino Acid Transport
Over the past three decades, a variety of mammalian amino acid transport systems
have been characterized extensively for cell and substrate specificity, kinetic parameters,
and metabolic regulation. However, the original studies began with the pioneering work
of Van Slyke and Meyer in 1913, when they demonstrated that tissues accumulated
amino acids against a concentration gradient (Van Slyke and Meyer, 1913). In the early
1960s, the Christensen laboratory began to define specific transport systems that mediate
the flux of amino acids across the membrane bilayer based on the conformation, size, and
chemical properties of the amino acid side chain (Oxender and Christensen, 1963).
Rigorous investigation revealed that each system recognizes more than one amino acid,
distinct systems exhibit some degree of overlapping substrate reactivity, and various
amino acids are transported by more than one system. These observations were further
complicated by the discovery that distinct cell and tissue types express a different
combination of systems that work in concert to provide specific nutritional requirements.
The identification of individual proteins responsible for amino acid transport activity has
been complicated by the relatively low abundance of these proteins, as well as the
technical difficulty involved in isolating and purifying integral membrane proteins.
However, the advances in molecular biology over the last ten years have resulted in the
1


128
Crdena, et al., 1997). Caveolin-bound eNOS is inactive until sufficient Ca2+-calmodulin
is present to compete for the caveolin binding site, thereby activating the enzyme to
produce NO. For previously unknown reasons, caveolar localization optimizes the ability
of eNOS to produce NO (Liu et al., 1996; Palmer and Moneada, 1989; Pollock et al.,
1991; Feron et al., 1996). The contribution of arginine transport to this phenomenon is
the subject of the research presented in this chapter.
An area under intense investigation concerns the mechanism by which specific,
yet unrelated, proteins are sequestered within the caveolae. In addition to sharing a
caveolin-binding motif, most of the proteins localized to caveolae are lipid modified.
There is significant evidence to suggest that acylation plays a role in targeting, although
the type and extent of the fatty acid requirement varies between different proteins. For
instance, eNOS is co-translationally myristoylated on glycine 2, and post-translationally
palmitoylated on cysteines 15 and 26. Fatty acylation moieties on these residues are
necessary, and sufficient, for targeting eNOS to caveolae (Garcia-Cardena et al., 1996).
The heterotrimeric G protein subunit az requires myristoylation at glycine 2 for stable
membrane association, and palmitoylation at cysteine 3 for specific localization to the
plasma membrane caveolae (Song et al., 1997; Morales et al., 1998). Myristoylation-
minus mutants of c-Src are poorly accumulated in caveolae, and other proteins, including
SNAP-25 and p59fyn, demonstrate a fatty acylation requirement for trafficking to the
plasma membrane, although caveolar localization has not been confirmed (Gonzalo and
Linder, 1998; van't Hof and Resh, 1997). However, in the case of p59fyn, it is believed
that protein synthesis and myristoylation occur on soluble ribosomes. This is followed by


75
abnormal cationic and neutral amino acid transport and accumulate abnormal intracellular
vesicles and vacuoles. Therefore, this project has focused on the evaluation of membrane
protein trafficking in the cells of LPI patients.
Trafficking Defects in Plasma Membrane Transport Proteins
There is precedence for the aberrant trafficking of a specific class of transporter
proteins in both Saccharomyces cerevisiae and Drosophila. It was shown that S.
cerevisiae requires an endoplasmic reticulum (ER) integral membrane protein, SHR3, for
the effective processing and trafficking of amino acid permeases, specifically (Ljungdahl
et al., 1992). Mutations in SHR3 cause retention of 13 out of 13 amino acid permeases
tested within the ER, and therefore, block amino acid uptake by interfering with the
plasma membrane localization. The defective SHR3 does not affect the targeting of any
other plasma membrane, secretory, or vacuolar protein (Ljungdahl et. ah, 1992). In
Drosophila, mutations in a gene (ninaA) encoding an ER cis-trans isomerase cause the
abnormal intracellular accumulation of two homologous opsin proteins in photoreceptor
cells. The accumulation is presumed to occur as a result of improper protein folding early
in the biosynthetic pathway (Colley et. ah, 1991).
In humans, a single amino acid deletion in a highly regulated chloride channel is
the basis for the fatal genetic disorder cystic fibrosis (CF). It was shown that the
mutation that was present in 70% of the defective CF genes (Kerem et ah, 1989) resulted
in the abnormal retention of the cystic fibrosis transmembrane conductance regulator
protein (CFTR) in the endoplasmic reticulum, although the CF conductance activity
measured in ER vesicles or reconstituted proteoliposomes was unimpaired (Cheng et ah,


2
cloning and expression of more than 20 cDNAs encoding amino acid transporters. Based
on sequence homology, a number of the cloned transporters have been classified
according to gene families. Those related to the studies presented in this thesis are
summarized in Table 1-1.
Table 1-1
cDNA Clones of Amino Acid Transporters in CAT and EAAT Families
Clone
Alternate
names
Deduced amino
acid length
Substrate
specificity
Ions coupled
CAT1

622-624
cationic

CAT2
CAT2a
658
cationic

CAT2a
CAT2b
659
cationic

CAT3

619
cationic

EAAT1
GLAST1
GluT
543
D,L-aspartate
L-glutamate
Na jn, K. out
H+in
EAAT2
GLT1, GLTR
GLAST2
573
D,L-aspartate
L-glutamate
Na+,, K+oul
H|n
EAAT3
EAAC1
523-525
D,L-aspartate
L-glutamate
Na+in, K+out
EAAT4
564
D,L-aspartate
L-glutamate
Na+in, K+out
H+in
EAAT5
560
D,L-aspartate
L-glutamate
Na+jn, K+oul
H,n
The "CAT family" is composed of four known transporters that mediate the Na+-
independent transport of the cationic amino acids arginine, lysine, ornithine, and histidine
when positively charged (reviewed by MacLeod et al., 1994; Malandro and Kilberg,
1996). Although the members of this family share a common substrate specificity and
significant amino acid similarity, they differ in tissue distribution, affinity for substrate,
and most likely the functional role they play in metabolism. The anionic transporter
family currently contains five members that are responsible for the Na+-dependent
transport of the anionic amino acids glutamate and aspartate (reviewed by Malandro and


4
(reviewed by Kilberg et al., 1993). Presently, four distinct proteins that specifically
mediate plasma membrane cationic amino acid transport have been cloned, and
demonstrate similar, yet not identical, properties to those of System y+.
I have chosen to study the primary cationic amino acid transporter, designated
CAT1, as well as several members of the glutamate family (described below). In 1989,
Cunningham and coworkers identified a cDNA clone that encoded the ecotropic murine
leukemia virus receptor (Albritton et al., 1989). It was later determined that the receptor
also functioned as a Na'-independent, high-affinity transporter for arginine, lysine,
ornithine, and histidine when positively charged (Kim et al., 1991; Wang et al., 1991).
The corresponding human cDNA was cloned by Meruelo and coworkers (Yoshimoto et
al., 1991) and is 88% homologous to the murine transporter, but lacks homology in the
region of viral binding. This is consistent with the observation that the ecotropic murine
leukemia retrovirus is unable to infect human cells (Albritton et al., 1993). The human
CAT1 sequence revealed a 629-amino acid protein with 12-14 predicted transmembrane
spanning domains. The CAT1 transporter mRNA is not expressed in the liver, but
otherwise appears to be ubiquitous (for review see Malandro and Kilberg, 1996).
Immunohistochemistry from our laboratory, using an antibody against the third
extracellular loop of the murine cDNA sequence, confirmed the lack of expression in rat
liver (Woodard et al., 1994).
MacLeod and coworkers cloned a second member of the CAT family (MacLeod
et al., 1990) from murine T cell lymphocytes (Kekuda et al., 1993). This protein,
originally called the Tea gene, was discovered during a search for genes involved in T
cell activation, however, its function as a transporter was suggested by its extensive


144
atherosclerosis, diabetes, hypercholesterolemia, hypertension, aging, cigarette smoking,
and heart failure (Harrison, 1996; reviewed by Harrison, 1997). Little is known regarding
the mechanisms of NO action underlying the above medical conditions. An alteration in
the expression of eNOS is one factor that may exacerbate a disease state. Although
eNOS is constitutively expressed, it is subject to regulation by vascular shear stress
(Nishida et al., 1992), exposure to lysophosphatidylcholine (Zembowicz et ah, 1995), low
concentrations of oxidized low density lipoprotein (Hirata et ah, 1995), and cyclic GMP
analogues (Ravichandran and Johns, 1995). During its life cycle, eNOS is dually-
acylated, phosphorylated at multiple sites, and interacts with caveolin and
calcium/calmodulin, if not several other proteins. A breakdown in one or more of these
processes would affect the production and release of NO. Other factors that may lead to
abnormal NO production include a change in the availability of cofactors, such as
tetrahydrobiopterin, or a destruction of NO by reactive oxygen species.
It is believed that the availability of the substrate, L-arginine, contributes to the
regulation of NO production, however conflicting data regarding the effectiveness of L-
arginine treatment exist. Mugge and Harrison reported, in isolated organ chambers, that
L-arginine had no effect on the vasodilation of aortas from normal and cholesterol-fed
rabbits (Mugge and Harrison, 1991). However, numerous studies have shown positive
vascular responses to L-arginine treatment in experimental animals and humans with
hypercholesterolemia, diabetes, and hypertension (Cooke et al., 1991; Creager et al.,
1992). Oral administration of L-arginine doubles the plasma levels of the amino acid and
has been effective in treating hypertensive rats, as well as, atherosclerosis in cholesterol-
fed rabbits (Cooke et al., 1992; Chen and Sanders, 1993). Although data concerning the


182
Shayakul, C., Kanai, Y., Lee, W.-S., Brown, D., Rothstein, J. D. and Hediger, M. A.
(1998) Am. J. Physiol. 273, F1023-F1029
Shupliakov, O., Brodin, L., Cullheim, S., Ottersen, O. P. and Storm-Mathisen, J. (1992)
J. Neurosci. 12, 3789-3803
Silbemagl, S. (1983) Pflugers Arch. 396,218-224
Simell, O. (1989) in The Metabolic Basis of Inherited Disease (Scriver, C. R., Beaudet,
A. L., Sly, W. S. and Valle, D., ed) pp. 2497-2513, McGraw-Hill, New York
Simell, O., Perheentupa, J., Rapla, J., Visakorpi, J. K. and Eskelin, L.- E. (1975) Am. J.
Med. 59, 229-240
Simionescu, M., Simionescu, N. (1987) in Pulmonary Endothelium in Health and
Disease (US Ryan, ed) pp. 35-62, Marcel Dekker, Inc., New York
Smith, D. W., Scriver, C. R., Tenenhouse, H. S. and Simell, O. (1987) Proc. Natl. Acad.
Sci. 84, 7711-7715
Smith, D. W., Scriver, C. R. and Simell, O. (1988) Hum. Genet. 80, 395-396
Solski, P. A., Quilliam, L. A., Coats, S. G., Der, C. J. and Buss, J. E. (1995) Meth.
Enzym. 250, 435-444
Song, K. S., Li, S., Okamoto, T., Quilliam, L. A., Sargiacomo, M. and Lisanti, M. P.
(1996) J. Biol. Chem. 271, 9690-9697
Stachowiak, M. K., Maher, P. A., Joy, A., Mordechai, E. and Stachowiak, E. (1996) Mol.
Cell Biol. 7,1299-1317
Steams, T. (1995) Curr. Biol. 5,262-264
Storck, T., Schulte, S., Hofmann, K. and Stoffel, W. (1992) Proc. Natl. Acad. Sci. U.S.A.
89, 10955-10959
Storm-Mathisen, J., Danbolt, N. C., Rothe, F., Torp, R., Zhang, N., Aas, J. E., Kanner, B.
L, Langmoen, I. and Ottersen, O. P. (1992) Prog. Brain Res. 94, 225-241
Szatkowski, M. and Attwell, D. (1994) Trends Neurosci. 17, 359-365
Tamarappoo, B. K., Handlogten, M. E., Laine, R. O., Serrano, M. A., Dugan, J., and
Kilberg, M. S. (1992) J. Biol. Chem. 267, 2370-2374
Tamarappoo, B. K., McDonald, K. K. and Kilberg, M. S. (1996) Biochim. Biophys. Acta
1219, 131-136




Figure 4-3. Intracellular staining of normal and LPI human fibroblasts with the CAT1
antibody. According to the procedures described in the Methods Chapter, normal (A) and
LPI (B) fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with an antibody specific for CAT1. The CAT1 antibody was
visualized by an FITC-labeled goat anti-rabbit IgG secondary antibody. Staining from
three independent experiments was analyzed by deconvolution microscopy and shown to
be reproducible. An outline was drawn around one cell to distinguish the periphery. The
data shown represent analysis of 0.2 pm sections through the cells.


99


ACKNOWLEDGMENTS
I would like to thank the members of my supervisory committee: Drs. Brian Cain,
William Dunn, Susan Frost, Mike Kilberg, and Peter McGuire. I wish to extend special
thanks to my mentor, Mike Kilberg, as well as Dr. Edward Block for his contribution to
the work discussed in Chapter 5. 1 would also like to acknowledge Dr. Stephen Wang for
his valuable instruction on the deconvolution microscope, David Parks, in the Center for
Structure Biology Computer Core, and Stephen Nowicki for valuable support and
friendship. Lastly, I would like to thank my parents, Drs. Maurice and Patricia
McDonald for their guidance, encouragement, and love.
li


40
Preliminary experiments with antibodies specific for EAAT2 and EAAT4 revealed no
detectable staining (data not shown). These experiments were intended to provide some
information concerning the similarities and differences in cellular distribution of amino
acid transporters belonging to the same gene family, and perhaps gain insight into why
multiple transporters with nearly identical kinetics are expressed in the same cell. For all
extracellular labeling experiments, fibroblasts were fixed with a 4% paraformaldehyde
solution and stained according to the immunofluorescence protocol described in the
Methods Chapter.
A 1:200 dilution of the anti-EAAT3 antibody stained the cells in clusters rather
than diffusely labeling the entire cell surface (Figure 3-1 A). This pattern of transporter
"patching" was similar to the pattern observed using the CAT1 arginine transporter
antibody (described in Chapter 5. Figure 5-1 A and B). Labeling by the anti-EAAT3
transporter antibody was completely inhibited by preadsorption of the antibody with 50
pg/ml of the corresponding peptide antigen for 12 h at 4C (Figure 3-1 B). The
intracellular distribution of the EAAT3 glutamate transporter was examined in human
fibroblasts following either -20C MeOH fixation, or 4% paraformaldehyde fixation in
combination with 0.1% Triton X-100 membrane permeabilization. Under both
conditions, the EAAT3 antibody labeled small vesicles throughout the cytoplasm (Figure
3-2 A). Double-labeling MeOH-fixed fibroblasts with antibodies against EAAT3 (1:200
dilution) and KDEL (1:100 dilution), a common epitope of resident proteins of the
endoplasmic reticulum (ER), showed very little co-localization (Figure 3-2 B). Also,
only a small amount of co-localization was detected when MeOFi-fixed fibroblasts were


74
observations in LPI patients, it was originally hypothesized that LPI may arise from a
defect in the cationic amino acid transporter, CAT1, previously termed System y+ (Smith
et al., 1987). However, two independent laboratories sequenced the human CAT1
mRNA expressed in LPI patients and found no mutations within the sequence (personal
communication, Dr. Olli Simell, University of Turku). As a result, our laboratory
hypothesized that a trafficking defect involving the CAT1 transporter, or a protein
involved in membrane protein trafficking, may be responsible for the altered transport of
cationic amino acids in LPI cells. Preliminary immunofluorescence studies in our
laboratory revealed a population of intracellular vesicles, unique to LPI cells, which
appeared to contain the CAT1 transporter (Woodard and Kilberg, unpublished data). The
elevated steady-state accumulation of lysine, described above, supports the hypothesis
that amino acids may be sequestered in the abnormal intracellular vesicles of LPI cells. If
the defect involves a deficiency in efflux across the basolateral membrane domain in
epithelial cells, as proposed, perhaps the reason is that the lysine becomes trapped within
the intracellular vesicles instead of rapidly equilibrating across the plasma membrane.
This hypothesis is consistent with the finding by Smith et al. that trans-stimulation of
cationic amino acid efflux is also decreased in LPI cells (Smith et al., 1987).
Dr. Olli Simell, at the University of Turku (Turku, Finland), has been following
Finnish patients with LPI for over 25 years. His laboratory has thoroughly documented
the clinical aspects of the disease and is currently working to localize and characterize the
defective gene. Using linkage analysis and a candidate gene approach, Simell and
coworkers have excluded the possibility that CAT1 or CAT2 (Lauteala et al., 1997) is the
defective gene. Despite these observations, it is clear that the LPI fibroblasts exhibit


70
(Simell. 1989). The concentrations of cationic amino acids in plasma are low. whereas
glutamine, alanine, serine, proline, citrulline, and glycine are increased. Although
cationic amino acid transport is probably defective in most cell types, it has been
documented to be so in kidney tubules, intestine, and cultured fibroblasts (Simell, 1989).
LPI-derived fibroblasts accumulate elevated steady state levels of cationic amino acids
and exhibit a reduced rate of trans-membrane exchange for these same substrates (Smith
et al., 1987). The biochemical basis for this decreased release of cellular amino acids has
not been established. Although several amino acidurias are thought to result from
defective NaT-dependent transport across the brush border domain of either intestinal or
renal epithelial cells, previous studies have suggested that the primary transport defect in
LPI is a reduced efflux across the basolateral surface (Simell, 1989; Rajantie et ah. 1981).
Evidence for this interpretation comes from an observation that plasma amino acid
concentrations remain low following the oral administration of both cationic amino acids
and lysine dipeptides. Dipeptides are transported normally across the brush border
membrane by a mechanism distinct from that of free amino acids. The dipeptides are
then hydrolyzed to free amino acids by intracellular enzymes. However, in LPI
enterocytes, these free amino acids are unable to pass through the basolateral membrane
into the plasma, and instead, are transported back out via a brush border membrane
transporter. Direct measurements of lysine transport in intestinal biopsy specimens have
confirmed that the transport defect is located on the basolateral membrane (Simell, 1989).
Transport appears to be normal at the luminal surface of renal epithelial cells from LPI
patients (Simell, 1989).


42
the other hand, very little co-localization was detected when EAAT1-R and KDEL
antibodies were used (data not shown), indicating that the fluorescence was not likely due
to staining of ER components surrounding the nucleus. A D77 antibody, generated
against the yeast nucleolar protein, fibrillarin (Noplp), was used at a dilution of 1:50 in
double-labeling experiments with EAAT1-R antibody (1:50 dilution) to determine if the
nuclear staining was nucleolar (Aris and Blobel. 1988). The D77 antibody detected three
or four nucleoli in each cell, however, no co-localization with EAAT1-R was observed
(Figure 3-3 D).
The nuclear membrane staining of human fibroblasts with an antibody against a
known plasma membrane amino acid transporter (EAAT1) was unexpected. In an effort
to confirm the nuclear localization of the EAAT1-R antibody, a second EAAT1 antibody
was obtained from Dr. Wilhelm Stoffel (University of Cologne. Germany) and will be
referred to as EAAT1-S (Table 3-1). Although EAAT1-S was also generated against an
N-terminal peptide, there was no overlap with the amino acid sequence that EAAT1-R
was raised against. Therefore, EAAT1-S provided an additional antibody in case the
EAAT1-R antibody was cross-reacting to an unknown protein with homology to the
peptide used to generate EAAT1-R. When MeOH-fixed human fibroblasts were
incubated with a 1:50 dilution of the anti-EAATl-S antibody, no nuclear membrane or
matrix staining was observed (Figure 3-4 A). Only small vesicles throughout the
cytoplasm were detected.
With the apparent contradicting results obtained from immunofluorescence
experiments using the EAAT1-R and EAAT1-S antibodies, two additional EAAT1


177
Gold, M. E., Bush, P. A., and Ignarro, L J. (1989) Biochem. Biophys. Res.Comunm. 164,
714-721
Greene, B., Pacitti, A. J., Souba, W. W. (1993) Am. J. Physiol. 264, L351-L356
Hanson, S. H., Sandvig, K. and van Deurs, B. (1993) J. Cell Biol. 123, 89-97
Harrison, D. G. (1996) Cardiol. Clin. 14, 1-15
Harrison, D. G. (1997) J. Clin. Invest. 100,2153-2157
Haugeto, O., Ullensvang, K., Levy, L., M., Chaudhry, F. A., Honor, T., Nielsen, M.,
Lehre, K. P. and Danbolt, N. C. (1996) J. Biol. Chem. 271, 27715-27722
Hecker, M., Sessa, W. C., Harris, H. J., Anggard, E. E., and Vane, J. R. (1990) Proc.
Natl. Acad. Sci. USA 87, 8612-8616
Hemler, M. E., Sanchez-Madrid, F., Flotte, T. J., Krensky, A. M., Burakoff, S. J., Bhan,
A. K., Springer, T. A. and Strominger, J. L. (1984) J. Immunol. 132, 3011-3018
Hershko, A. and Ciechanover, A. (1992) Annu. Rev. Biochem. 61, 761-807
Hiraoka, Y., Sedat, J. W., and Agard, D. A. (1987) Science 238, 36-41
Hiraoka, Y., Swedlow, J. R., Paddy, M. R., Agard, D. A. and Sedat, J. W. (1991) Sem.
Cell Biol. 2, 153-165
Hirata, K., Miki, N., Kuroda, Y., Sakoda, T., Kawashima, S. and Yokoyama, M. (1995)
Circ. Res. 76, 958-962
Hosokawa, H., Sawamura, T., Kobayashi, S., Ninomiya, H., Miwa, S. and Masaki, T.
(1997) J. Biol. Chem. 272,8717-8722
Ito, K. and Groudine, M. (1997) J. Biol. Chem. 272, 26780-26786
Ivessa, N. E., De Lemos-Chiarandini, C., Gravotta, D., Sabatini, D. D. and Kreibich, G.
(1995) J. Biol. Chem. 270,25960-25967
Jones, K. L., Stewart, R. M., Fowler, M., Fukuda, M. and Holcombe, R. F. (1992) Clin.
Immunol. Immunopathol. 65, 219-226
Juliano, R. L. and Ling, V. (1976) Biochim. Biophys. Acta 455, 152-162
Kakuda, D. K., Finley, K. D., Dionne, V. E. and MacLeod, C. L. (1993) Transgene 1,
91-101


53
10.0 um


5
sequence homology to CAT1. The Tea gene, later renamed murine CAT2, is 61%
identical to CAT1 at the amino acid level, has 12 predicted membrane spanning domains,
and mediates Na'-independent high-affinity cationic amino acid transport in activated T
lymphocytes (for review see Malandro and Kilberg, 1996). Chimeric constructs of the
mouse and human CAT1 sequences were used to identify the region of viral binding in
the third extracellular loop of the murine CAT1 protein (Albritton et al., 1993). The
CAT2 sequence, like the human CAT1 sequence, is divergent in this region and does not
function as a binding site for the retrovirus. Nitric oxide (NO) has been implicated in T
cell signaling by autocrine/paracrine pathways, and it has been proposed that the
expression of the CAT2 transporter during T cell activation may be related to the need for
arginine in the production of NO (MacLeod et al., 1994). Using the murine CAT2
sequence as a probe for screening a mouse liver cDNA library, Cunningham and
coworkers identified a third liver-specific member of the CAT family, murine CAT2a.
This protein is encoded by the same gene as CAT2, but as a result of differential splicing,
CAT2a has an additional stretch of 41 amino acid residues (358-398) between the eighth
and ninth membrane spanning domains. Despite the similarity in sequence, CAT2a
exhibits a 10-fold lower affinity for arginine than either CAT1 or CAT2 (Closs et al.,
1993). This kinetic difference suggests that the 41 amino acid region may be involved in
binding the amino acid during its translocation across the membrane.
Screening a rat brain cDNA library with probes designed from the murine CAT1
sequence isolated the most recent addition to the CAT family, CAT3 (Hosokawa et al.,
1997). The protein encoded by the rat CAT3 is comprised of 619 amino acids and shares
53-58% identity with the CAT family members previously described. The CAT3 protein


48
suggest that the EAAT1-S, but not the EAAT1-R, antibody recognizes the expressed
EAAT1-GFP(N3) protein in the human fibroblasts.
Co-localization of EAAT1-GFP(N3) with the EAAT1 endogenous in PAEC. As
an extension of the experiments in the last section. PAEC were transfected with the
EAAT1-GFP(N3). fixed with -20C MeOH. and stained with either the EAAT1-R (Figure
3-9 A) or EAAT1-S (Figure 3-9 B) antibody. Primary antibodies were detected using a
1:200 dilution of goat anti-rabbit IgG linked to Texas Red. PAEC were selected because
they were the only cell type tested that did not show nuclear localization of EAAT1 in the
initial experiments with the EAAT1-R antibody. As observed with the human
fibroblasts, there was significant co-localization between the EAAT1-GFP(N3) fusion
protein and the EAAT1-S antibody in PAEC. Although the fluorescent patterns of the
EAAT1-R antibody and EAAT1-GFP(N3) were both punctate and cytoplasmic, there was
little overlap between the two stains.
Discussion
This project began with the observation that an antibody (EAAT1-R) against a
known plasma membrane transporter (EAAT1) labeled the nucleus of human fibroblasts.
The EAAT3 isoform, on the other hand, was detected on the plasma membrane and in
vesicles throughout the cytoplasm. The clustering observed using the anti-EAAT3
antibody is consistent with the report by Davis et al. showing punctate fluorescence
throughout the cytoplasm with clustering at the cell surface (Davis et al., 1998).
Although the EAAT1 nuclear localization was unexpected, it was not unprecedented. As
discussed in the Introduction to this chapter, Matthews et al. detected nuclear staining


164


183
Tanaka, K., Watase, K., Manabe, T., Yamada, K., Watanabe, M., Takahashi, K., Iwama.
H., Nishikawa, T., Ichihara, N., Kikuchi, T., Okuyama, S., Kawashima, N., Hori,
S. and Takimoto, M. (1997) Science 276, 1699-1702
Taylor, P. D.and Poston, L. (1994) Br. J. Pharmacol. 113, 801-808
Teitel, J. M. (1986) J. Biol. Chem. 128,329-336
Utsunomiya-Tate, N., Endou, H. and Kanai, Y. (1996) J. Biol. Chem. 271, 14883-14890
Vadgama, J. V. and Christensen, H. N. (1984) J. Biol. Chem. 259, 3648-3652
Van Aelst, L., Barr, M., Marcus, S., Polverino, A. and Wigler, M. (1993) 90 Proc. Natl.
Acad. Sel, U. S. A.. 90, 6213-6217
Van Slyke, D. D. and Meyer, G. M. (1913) J. Biol. Chem. 16, 197
Van Winkle, L. J. (1988) Biochim. Biophys. Acta 947, 173-208
Van Winkle, L. J., Campione, A. L. and Gorman, M. L. (1988) J. Biol. Chem. 263,
3150-3163
Van Winkle, L. J., Christensen, H. N. and Campione, A. L. (1985) J. Biol. Chem. 260,
12118-12123
Velaz-Faircloth, M., McGraw, T. S., Malandro, M. S., Fremeau, R. T., Kilberg, M. S.,
and Anderson, K. J. (1996) Am. J. Physiol. 270, 67-75
Wahle, S. and Stoffel, W. (1996) J. Cell Biol. 135, 1867-1877
Wang, H., Kavanaugh, M. P., North, R. A., and Rabat, D. (1991) Nature 352, 729-731
Wang, Y. and Tate, S. S. (1995) FEBSLett. 368, 389-392
Ward, D. M Leslie, J. D. and Kaplan, J. (1997) J. Cell Biol. 139, 665-673
Watts, C. and Marsh, M. (1992) J. Cell Sci. 103, 1-8
Wells, R. G., Lee, W., Kanai, Y., Leiden, J. M. and Hediger, M. A. (1992) J. Biol. Chem.
267, 15285-15288
White, M. F. (1985) Biochim. Biophys. Acta 822, 355-374
White, M. F. and Christensen, H. N. (1983) J. Biol. Chem. 258, 8028-8038
White, M. F., Gazzola, G. C and Christensen, H. N. (1992) J. Biol. Chem. 257, 4443-
4449


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E8XRUX60C_N2M8Q3 INGEST_TIME 2014-10-07T21:00:14Z PACKAGE AA00025807_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


55


147
labile free radical that must reach the target tissue within approximately 10 seconds. The
CATl-eNOS caveolar machinery would provide the necessary degree of efficiency and
regulation for proper NO signaling. If this model holds, it would represent the first
example, in mammalian cells, of a functional complex between an amino acid transport
protein and an enzyme. Although a direct interaction has not been proven,
immunofluorescence and vesicle transport data indicate that a close spatial alignment and
functional relationship exist between CAT1 and eNOS in the caveolae. Directed delivery
of extracellular arginine to eNOS would account for the arginine paradox described
earlier (Kurz and Harrison, 1997) and would also explain the observation by Liu et al.
that caveolar localization of eNOS is required for optimal NO production by eNOS (Liu
et al., 1996). Understanding the subcellular environment and events that are required for
NO production in endothelial cells will provide a better understanding of the mechanisms
that regulate blood pressure in the pulmonary circulation of patients with lung disease.


80
transporter staining to disperse over the entire cell surface, and removal of the inhibitor
caused the clusters of transporter to reform within 3 h (data discussed in Chapter 5).
These data confirmed previous reports by Woodard et al. that the CAT1 antibody forms
clusters on the plasma membrane of normal fibroblasts, and that this arrangement is
dependent on intact microtubules (Woodard et al., 1994). Identification of these clusters
as caveolae will be discussed in Chapter 5.
In a separate series of experiments, the intracellular distribution of the CAT1
transporter was detected by immunofluorescence and deconvolution microscopy in
MeOH-fixed fibroblasts. Although an intracellular vesicle population was labeled with a
1:25 dilution of the CAT1 antibody in both normal (Figure 4-3 A) and LPI fibroblasts
(Figure 4-3 B), there was not a significant difference in the number or physical
appearance of the vesicles. Incubating cells with secondary antibody only, or incubating
the primary antibody with 50 pg/ml of corresponding peptide, resulted in only a faint,
diffuse background immunofluorescence (data not shown). These results are different
than earlier reports from our laboratory that documented an obvious difference between
the normal and LPI cells when observed using an epifluorescence microscope. It was
observed previously that CAT1 antibody labeling of normal fibroblasts resulted in diffuse
cytoplasmic staining with slightly increased intensity in the region of the Golgi complex
(a perinuclear pattern). On the other hand, the CAT1 antibody staining in the LPI cells
was specific for a large number of intracellular vesicles, which appeared to be randomly
distributed throughout the cytoplasm. As a result of the apparent discrepancy, staining of
normal and LPI fibroblasts was repeated using a variety of conditions. CAT1 serum.


51
membrane markers need to be used to determine whether or not the EAAT1
immunoreactivity is a result of the contamination of nuclear preparations with other
membranes. Future experiments could also include the isolation of the EAAT1
immunoreactive band from an SDS-PAGE gel for protein sequencing, as well as
immunofluorescence on isolated nuclear fractions.


136
expression resulted in large quantities of GFP(C3)-CAT1 on the plasma membrane, there
were still specific regions of more highly concentrated transporter. The CAT1 antibody
(1:50 dilution) showed strong co-localization with the expressed GFP(C3)-CAT1 in
PAEC, as illustrated in Figure 5-5 B. However, there were separate pools of unstained
GFP(C3)-CAT1 fusion protein and CAT1 antibody (detected by a 1:200 dilution of goat
anti-rabbit IgG linked to Texas Red). This was probably due to CAT1 antibody staining
of endogenous transporter, as well as a less than 100% efficiency of CAT1 antibody
staining of the expressed GFP(C3)-CAT1 protein. To determine if the GFP(C3)-CAT1
co-localized with eNOS in the caveolae, PAEC were transfected with the fusion protein,
fixed with -20C MeOH, and stained with caveolin (Figure 5-6 A) or eNOS (Figure 5-6
B) antibodies. Both antibodies were visualized using goat anti-mouse IgG conjugated to
Texas Red. Although significant co-localization of GFP(C3)-CAT1 with caveolin or
eNOS existed, the over-expression also resulted in a greater quantity of transporter that
did not appear to exist in caveolae.
Co-localization of CAT1 and eNOS following nocodazole treatment. As
mentioned previously, the CAT1 clusters can be dispersed by treatment with the
microtubule-destabilizing drug, nocodazole (Woodard et al., 1994). Several reports have
suggested that caveolin also relies partially on microtubules for cycling between the
Golgi and the plasma membrane caveolae (Conrad et al., 1995). To determine whether or
not CAT1 and eNOS are dependent on microtubules for co-localization, PAEC were
treated with MEM + 25 pg/ml nocodazole (prepared from a 3mg/ml nocodazole stock in
DMSO) for 1 h at 37C before double-staining with anti-CATl and anti-eNOS antibodies


CHAPTER 4
LYSINURIC PROTEIN INTOLERANCE
Introduction
Lysinuric Protein Intolerance (LPI) is an autosomal recessive disease that is
characterized by a defect in dibasic amino acid transport, as well as, an impaired urea
cycle (reviewed by Simell, 1989). Although cases have been reported worldwide, the
highest prevalence is in Finland, where LPI afflicts 1 in 60,000 to 80.000 people.
Patients have an extremely low tolerance for dietary protein, and symptoms of
hyperammonemia are revealed shortly after weaning infants from the high-fat, low-
protein breast milk. Nausea, vomiting, and diarrhea following meals are early indications
of the disorder. Throughout infancy and childhood, patients show signs of growth
retardation and fail to thrive as a result of protein deficiency. They have enlarged livers
and spleens, muscle hypotonia and hypertrophy, osteoporosis, and approximately 20% of
the patients show varying degrees of mental retardation. The only available treatment is
to limit the consumption of protein. Normalization of hepatic nitrogen utilization and
urea synthesis requires supplementing meals with 3 to 8 grams of citrulline daily.
LPI was originally defined by elevated urinary levels and poor intestinal
absorption of all cationic and several neutral amino acids (Simell, 1989). Normal daily
urine contains a mean of 4.13 mmol lysine/1.73 m2 body surface area, whereas the urine
from LPI patients contains a mean of 25.7 mmol lysine/1.73 m2 body surface area
69


19
passage. Incubation of all cell lines described was at 37C under a humidified atmosphere
of 5% C02-95% air.
Immunohistochemistrv. For immunofluorescence assays, cells were transferred to
22 x 22 mm sterilized Corning glass microscope cover slips, by placing the cover slips in
the wells of the Falcon six-well cluster trays, and plating the cells, which were then
allowed to reach 60 to 70% confluence. PAEC cells required pre-treatment of the cover
slips with 7-10 fig/ml flbronectin, as described above for culture dishes. Following three
5-min washes with phosphate buffered saline (PBS) to remove culture medium, cells
were fixed with a 4% paraformaldehyde solution for 20 min. To prepare the fixation
solution, 4% paraformaldehyde was added to 50% of the final volume of water and the
mixture was heated to 60-65C. Drop-wise addition of 10N sodium hydroxide (usually 1-
2 drops) was required to completely dissolve the paraformaldehyde. After the solution
had cooled, 30% of the final volume of water and 20% of the final volume of 5X PBS
were added, and the solution was adjusted to a pH of 7.5 to 8.0. After incubation with the
cells, the fixative was removed with three 5-min washes in PBS, and any residual
paraformaldehyde was blocked with 50 mM glycine (in PBS) for 30 min. This
incubation was followed by three additional 5-min PBS washes. Paraformaldehyde
fixation was used for immunofluorescence experiments designed to examine plasma
membrane labeling. If cell permeabilization was desired in combination with
paraformaldehyde fixation, 0.1 to 0.2% Triton X-100 was added to the wells for the last
30 minutes of blocking. Alternatively, a -20C methanol incubation for 5-min was used
to fix cells for the purpose of intracellular staining. Following either fixation procedure,


89
described in the Methods Chapter. After 24 h, the cells were fixed with -20C MeOH,
and stained with organelle-specific antibodies (according to normal transfection and
immunofluorescence procedures described in the Methods Chapter). The GFP-CAT1
transfection was used instead of the CAT1 antibody because many of the organelle-
specific antibodies were generated in the same species as the CAT1 antibody, and
therefore, could not be used for double-labeling experiments. The Rab5 antibody was
used to co-localize GFP-CAT1 with early endosomes, the mannose-6-phosphate receptor
antibody was used to detect GFP-CAT1 in Golgi and late endosomes/TGN, and the
lpgl 20 antibody was used to observe GFP-CAT1 associated with the late endosomes and
lysosomes. The results obtained were the same for experiments conducted in both normal
and LPI fibroblasts. The GFP-CAT1 co-localized to a moderate degree with Rab5 and
mannose-6-phosphate receptor antibodies (data not shown); however, very little overlap
was observed with the lpgl20 antibody (Figure 4-14 A and B). This was a reasonable
result given the probable participation of the transporter in the biosynthetic and recycling
pathways. Even though the lysosomes, or a lysosome-like vesicle population, appear to
be abnormal in the LPI cells, there was no accumulation of the GFP-CAT1 in these
compartments.
Discussion
Previous experiments demonstrated that LPI fibroblasts exhibited an elevated
accumulation of cationic amino acids over time, as well as the existence of a unique
vesicle population. As a result, it was hypothesized that the CAT1 transporter may be
trapped in an intracellular compartment and, therefore, unable to mediate efflux of


49
with the EAAT1-R antibody in the maternal decidua and placental trophoblast cells of rat
placenta using immunocytochemistry (Matthews et al.. 1998). In other cells of the
placenta, no nuclear localization was detected, but rather, EAAT1 was distributed
throughout the cytoplasm and on the plasma membrane. Other plasma membrane
proteins, such as P-glycoprotein and various growth factor receptors, have also been
detected in the nucleus and nuclear membrane by immunohistochemistry (Stachowiak et
al., 1996; Baldini et al 1995).
There are various explanations for the data presented above. First, the EAAT1
glutamate transporter may reside in the nucleus and provide a function that is different
from that of the plasma membrane nutrient transporter. If EAAT1 is serving as an amino
acid transporter in the nucleus, it is unlikely to be mechanistically similar to that on the
cell membrane due to the lack of a Na* gradient across the nuclear membrane. However,
the glutamate transporters also carry other ions such as, K* and H\ and the isoforms.
EAAT4 and EAAT5, have intrinsic chloride channel properties. Perhaps, ion transfer is
the primary function within the nuclear membrane. On the other hand, the lack of nuclear
localization observed with the EAAT1-GFP(N3) and GFP(C3)-EAAT1 fusion proteins,
as well as with both the EAAT1-S and EAAT1-D antibodies, presents evidence against
nuclear localization of the native EAAT1 protein. Although no nuclear localization
signaling sequence has been identified in EAAT1, there are hundreds of nuclear
localization sequence variations which are not highly conserved (Boulikas, 1996).
Another alternative is that the protein recognized by the EAAT1-R antibody
might prove to be a nuclear isoform of EAAT1 with a conserved sequence at the N-
terminus. The fact that the EAAT1-S antibody, generated against a different peptide


Figure 4-14. Lysosomal staining of normal and LPI cells expressing the GFP(C3)-CAT1
fusion protein. Normal (A) and LPI (B) fibroblasts were transfected for 3 h with the
GFP(C3)-CAT1 fusion protein according to the lipofectamine protocol described in the
Methods Chapter. Following 24-48 h of expression, cells were fixed with -20C MeOH
and stained with the anti-lpgl20 antibody. The lpgl20 primary antibody was detected
with a goat anti-rabbit IgG conjugated to Texas Red. Images were processed from three
independent transfections and the fluorescence patterns of expressed proteins were
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.


Figure 3-3. Nuclear staining of human fibroblasts with EAAT1-R antibody and co
localization with nucleus-specific antibodies. Using the methods described in the
Methods Chapter, human fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with antibodies specific for EAAT1-R (A), EAAT1-R that had
been preadsorption for 12 h at 4C with 50 pg/ml of the corresponding peptide antigen
(B), EAAT1-R and 414 (C), or EAAT1-R and D77 (D). Co-localization of the proteins
was assayed by using simultaneously a rabbit polyclonal antibody against EAAT1-R
detected by FITC-labeled goat anti-rabbit IgG, and mouse monoclonal antibodies against
414 and D77 detected by Texas Red labeled goat anti-mouse IgG. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. An outline was drawn around one cell in each panel to distinguish the
periphery. The data shown represent analysis of 0.2 pm sections through the cells.


145
responses to L-arginine have been collected for a number of diseases, in a variety of
experimental models, the mechanisms underlying the results remain elusive. L-arginine
may work directly on the endothelium through NO as a vasodilator, or it may block the
actions of endogenous NOS antagonists such as asymmetric dimethyl arginine (ADMA)
(Boger et al., 1997). This chapter presents evidence to support the concept that a
functional relationship, between eNOS and the CAT1 arginine transporter, provides
several additional potential regulatory mechanisms for the production of NO.
Immunofluorescence studies to detect endogenous proteins, as well as transfection
of a GFP(C3)-CAT1 construct, showed that caveolin, CAT1, and eNOS co-localized to
specific regions within the PAEC. In addition, an eNOS-specific antibody was able to
immunoprecipitate CAT 1-mediated transport activity. These data document a close
spatial alignment, and even suggest but do not prove a direct binding between these
proteins. Several laboratories have shown a direct interaction between caveolin-1 and
eNOS by reciprocal immunoprecipitations and yeast two-hybrid systems. However, our
attempts to immunoprecipitate eNOS with an antibody against CAT1 were unsuccessful,
even after cross-linking reagents were used to prevent dissociation during somewhat
harsh isolation conditions. There are numerous possible reasons for these negative
results. It is possible that the CAT1 antibody, although able to detect the transporter
during immunostaining protocols, was unable to recognize or interact with CAT1 bound
to eNOS following solubilization of the complex. The results from the PAEC
reconstituted vesicle transport following immunodepletion provide evidence for a direct
interaction between CAT1 and eNOS. The attempt to detect eNOS on an immunoblot
after immunoprecipitating with the CAT1 antibody would have provided additional


Figure 4-4. Expression of GFP and the GFP(C3)-CAT1 fusion protein in normal human
fibroblasts. Normal fibroblasts were transfected for 3 h with GFP(C3) only (A), or with
the GFP(C3)-CAT1 fusion protein (B) according to the lipofectamine protocol described
in the Methods Chapter. Following 24-48 h of expression, cells were fixed with -20C
MeOH and visualized by deconvolution microscopy. In Panel C, GFP(C3)-CAT1
transfected cells were stained with an antibody against CAT1 and detected with a Texas
Red-labeled goat anti-rabbit IgG. Images were processed from three independent
transfections and the fluorescence patterns of expressed proteins were determined to be
reproducible. The data shown represent analysis of 0.2 pm sections through the cells.


86
in the structure of the Golgi, or any component of the biosynthetic pathway that was
investigated in the normal and LPI fibroblasts.
Organelles of the recycling pathway were visualized using a 1:100 dilution of
rabbit anti-human Rab5 antibody (Figure 4-9 A and B), for labeling early endosomes, and
a 1:5 dilution of mouse anti-human transferrin receptor antibody (Figure 4-9 C and D),
for detecting vesicles involved in endocytosis and recycling. Neither of these antibodies
detected an abnormality in the recycling pathway of the LPI cells. The number of
vesicles containing the transferrin receptor varied significantly between experiments, but
the variation was independent of normal versus LPI. Vesicles and compartments
involved in degradation were identified using a 1:300 dilution of the lysosomal protease,
cathepsin D (Figure 4-10 A and B). Although the antibody demonstrated a punctate
staining pattern in both cell types, the vesicles detected in the LPI fibroblasts (Figure 4-10
B) were larger and in greater abundance than those observed in the normal fibroblasts
(Figure 4-10 A). To further test for a difference in the lysosomal staining, normal and
LPI cells were labeled with a 1:100 dilution of a mouse antibody generated against the
lysosomal membrane protein, lpgl 20 (Figure 4-11 A and B). In the normal fibroblasts
(Figure 4-11 A), a punctate pattern of lysosomal labeling was observed throughout the
cytoplasm. On the other hand, the apparent diameter of the lysosomes detected in the LPI
fibroblasts (Figure 4-11 B) were larger and in greater abundance than in the normal cells.
In addition, the LPI lysosomes were tightly clustered and located in close proximity to the
nucleus.
Treatment of normal and LPI fibroblasts with lysosomotropic agents. From the
labeling, it appeared as though the lpg!20-containing lysosomes corresponded to the


38
0.5-1.0 pg/pl in sample dilution buffer and 10-20 pg protein per lane were subjected to
one-dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE)
(Laemmli. 1970; Chiles et al., 1987). The proteins were transferred at 299 mAmps for 18
h to a nitrocellulose membrane (Chiles et al., 1987). The nitrocellulose membrane was
blocked with 5% non-fat dry milk (NFDM) at room temperature for 1.5 h. rinsed with
TBS/TWEEN (30 mM Tris base, 150 mM NaCl, 0.1% Tween 20, pH 7.6), then incubated
with primary antibody (summarized below) prepared in TBS/TWEEN for 1-2 h at room
temperature. Two quick rinses, one 15 min rinse, and two 5 min rinses with
TBS/TWEEN were followed by incubation of the nitrocellulose in the secondary
antibody, prepared in TBS/TWEEN, for 1-2 h at room temperature. The EAAT1-R
antibody was used at 1:200-1:1000 dilutions, the EAAT1-S antibody was used at a 1:50
dilution, and both were visualized using a 1:5000 dilution of either a donkey or goat anti
rabbit IgG conjugated to horseradish peroxidase. The EAAT1-C antibody was used at a
dilution of 1:100 and detected with a 1:10,000 dilution of goat anti-guinea pig IgG
conjugated to horseradish peroxidase. After washing the nitrocellulose blot six times for
5 min each with TBS/TWEEN, the blot was incubated in 6 ml of a 1:1 mixture of the
Enhanced Chemiluminescence (ECL) reagents (Pierce, Rockford. IL) for 1 min, drained,
wrapped in plastic, and immediately exposed to film.
Antibodies for immunoblotting and immunofluorescence. The antibodies used for
immunoblotting (described above) and immunofluorescence (described in Chapter 2) are
summarized in Table 3-1. Unless otherwise stated, EAAT1 and EAAT3 antibodies were
detected during immunofluorescence assays using a 1:200 dilution of goat anti-rabbit IgG


22
time variations. Following image correction, 3-D deconvolution corrected for the out-of-
focus contamination from each optical section. The images were displayed using an
integrated, multiple-windowed, mouse-driven display and Delta Vision software (Applied
Precision. Issaquah, WA).
Expression of exogenous transporter by transfection. The conditions for
transfection of the transporter cDNAs were optimized using human fibroblasts. The same
transfection protocol was used for Hela, PAEC, and HepG2 cells. Cells were plated onto
cover slips in 6-well trays 24 h before transfection. Optimal density for transfection was
within the range of 60-80% confluence for all cell lines used. Cells were transfected
using lipofectamine in Opti-MEM I reduced-serum medium according to the standard
protocol from Gibco Laboratories. For each transfection, 1 pg of a cDNA was added to
100 pi of Opti-MEM in a 17 x 100 mm polystyrene tube, and 4 pi of lipofectamine
reagent was mixed with 100 pi of Opti-MEM medium in a second polystyrene tube. The
solutions in each tube were combined and allowed to incubate for 30 min at room
temperature in order for cDNA-liposome complexes to form. During the incubation, cells
were washed two times quickly with PBS. For each transfection, 0.9 ml of serum- and
antibiotic-free MEM medium was mixed with the cDNA-liposome complexes and 1 ml
of the final transfection solution was applied to the cells. The cells were incubated with
the mixture for 3 h at 37C before washing two times with PBS and adding MEM
supplemented with antibiotics and 10% FBS. After 24 h, cells were fixed and labeled
according to the immunofluorescence protocol as described above. This lipofectamine
protocol above was compared to the liposome-mediated transfection protocols from either


Figure 5-11. eNOS staining of PAEC treated with varying concentrations of extracellular
L-arginine. PAEC were incubated with MEM + 10% FBS (A), MEM minus L-arginine
+ 10% FBS (B), and MEM + 0.5 mM L-arginine + 10% FBS (C) for 12 h under normal
culture conditions (see Methods Chapter). Cells were then fixed with -20C MeOH and
stained with anti-eNOS antibody detected by goat anti-mouse IgG linked to FITC.
Images were processed from three independent experiments and the staining was
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.


Figure 4-10. Intracellular staining of normal and LPI fibroblasts with an antibody against
a lysosomal enzyme. According to the protocol described in the Methods Chapter,
normal (A) and LPI (B) fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with the anti-cathepsin D antibody specific for lysosomes and
related acidic vesicles. The anti-cathepsin D primary antibody was visualized with a goat
anti-rabbit IgG conjugated to FITC. Staining from three independent experiments was
analyzed by deconvolution microscopy and shown to be reproducible. Only one cell is
shown in panel A, whereas one of several cells is outlined in panel B in order to
distinguish the cell periphery. The data shown represent analysis of 0.2 pm sections
through the cells.


179
MacLeod, C. L., Finley, K. D., Kakuda, D.K. (1994) J. Exp. Biol. 196, 109-21
MacLeod, C. L., Finley, K., Kakuda, D., Kozak, C. A. and Wilkinson, M. F. (1990) Mol.
Cell. Biol. 10, 3663-3674
Makowske, M. and Christensen, H. N. (1989) J. Biol. Chem. 257, 5663-5670
Malandro, M. S. and Kilberg, M. S. (1996) Annu. Rev. Biochem. 65, 305-336
Mancini, G. M. S., Beerens, E. M. T., Aula, P. P. and Verheijen, F. W. (1991) J Clin.
Invest. 87, 1329-1335
Manfras, B. J., Rudert, W. A., Trueco, M. and Boehm, B. O. (1994) Biochim. Biophys.
Acta, 1195, 185-188
Matthews, J. C., Beveridge, M. J., Malandro, M. S., Rothstein, J. D., Campbell-
Thompson, M., Verlander, J. W., Kilberg, M. S. and Novak, D. A. (1998) Am. J.
Physiol. 274, C603-C614
Mattson, M. P., Lee, R. E., Adams, M. E., Gutherie, P. B. and Kater, S. B. (1988) 1
Neuron 1, 865-876
McDonald, K. K., Rouhani, R., Handlogten, M.E., Block, E. R., Griffith, O. W., Allison,
R. D and Kilberg, M. S. (1997) Biochim. Biophys. Acta 1324, 133-141
McDonald, K. K Zharikov, S Block, E. R and Kilberg, M. S. (1997) J. Biol. Chem.
272,31213-31216
Michel, J., Feron, 0., Sacks, D. and Michel, T. (1997) J. Biol. Chem. 272, 15583-15586
Mitchell, J. A.,Flecker, M., Anggard, E. E., and Vane, J. R. (1990) Eur. J. Pharmacol.
182, 573-576
Moneada, S. and Higgs, A. (1993) N. Engl. J. Med. 329, 2002-2012
Moneada, S., Palmer, R. M., and Higgs, E. A. (1991) Pharmacol. Rev. 43, 109-142
Morin, A. M. and Stanboli, A. (1993) J. Neurosci. Res. 36, 272-279
Mugge, A. and Harrison, D. G. (1991) Blood Vessels 28, 354-357
Nathan, C. and Xie, Q. (1994) Cell 78, 915-918
Nicholls, D. G. (1993) Eur. J. Biochem. 212, 613-631
Nishida, K., Harrison, D. G., Navas, J. P., Fisher, A. A., Dockery, S. P., Uematsu, M.,
Nerem, R. M., Alexander, R. W. and Murphy, T. J. (1992) J. Clin. Invest. 90,
2092-2096


10
EAAT4 and EAAT5 have shown a large chloride conductance in addition to transport
activity. In the case of EAAT5, this associated chloride conductance may participate in
visual processing (Arriza et al., 1997). A potential PSD-95-binding motif was identified
in the C-terminus of EAAT5 (Arriza et al.. 1997). PSD-95, a cytoskeleton-associated
synaptic protein, has been shown to bind to C-terminal sequences in both the N-methyl-
D-aspartate (NMDA) receptor and Shaker-type potassium channels (Cho et al., 1992).
ASCT Amino Acid Transport Systems
Two zwitterionic amino acid transporters, ASCT1 (Shafqat et al., 1993; Arriza et
al., 1993) and ASCT2 (Kekuda et al., 1996; Utsunomiya-Tate et al., 1996), are 56%
identical to one another at the amino acid level, and both are approximately 40% identical
to the five glutamate transporters. The "ASC" transporters are Na-dependent, and
although they exhibit a broad substrate specificity, they prefer amino acids with
hydroxyl- or sulfydryl-containing side chains (serine, cysteine, and threonine). Like
System ASC, the transporters exhibit the property of trans-stimulation even though this
activity is typically a characteristic of Na+-independent transporters. Northern blot
analysis revealed highest expression of ASCT1 in the brain, skeletal muscle, and pancreas
(Shafquat et al.. 1993; Arriza et al., 1993). ASCT2, also known as ATB, mRNA is
expressed in lung, skeletal muscle, kidney, large intestine, testes, and adipose tissue
(Kekuda et al., 1996; Utsunomiya et al., 1996). This transporter, only recently cloned by
homology to ASCT1, has been shown to exhibit a broader substrate specificity than
ASCT1 (Kekuda et al., 1996; Utsunomiya et al., 1996). In fact, one of the distinguishing


Ill


162
A B
* 140 kDa
Figure 5-8. CAT1 immunoprecipitation of eNOS from solubilized PAEC plasma
membrane vesicles. The CAT1 antibody was used to immunoprecipitate eNOS
according to the protocol in the Methods Section of Chapter 5. Lane A shows eNOS
pelleted by control IgG linked to protein A-sepharose beads, presumably through non
specific binding, and lane B shows eNOS immunoprecipitated with the protein A-
Sepharose beads conjugated to the CAT1 antibody. Immunoblot analysis was performed
with a 1:200 dilution of eNOS antibody detected by a 1:20,000 dilution of goat anti
mouse IgG conjugated to horseradish peroxidase. Data shown are representative of five
individual blots.


Figure 3-4. Intracellular staining of human fibroblasts with EAAT1-S and EAAT1-C
antibodies. Using the methods described in the Methods Chapter, human fibroblasts were
fixed with -20C MeOH and subjected to immunohistochemistry with antibodies specific
for EAAT1-S (A) and EAAT1-C (B). The EAAT1-S antibody was detected by an FITC-
labeled goat anti-rabbit IgG, and the EAAT1-C antibody was detected by a Cy3-labeled
goat anti-guinea pig IgG. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells.


83
with the GFP(C3)-CAT1 demonstrated specific targeting to the plasma membrane, as
well as one or more intracellular vesicle populations (Figure 4-4 B). The strong
perinuclear staining is probably GFP(C3)-CAT1 in the Golgi. The expressed GFP(C3)-
CAT1 fusion protein also showed significant co-localization with the CAT1 antibody
(detected with a 1:200 dilution of goat anti-rabbit IgG linked to Texas Red) in normal
fibroblasts (Figure 4-4 C). Although the GFP(C3)-CAT1 made it easier to visualize the
exact location of the CAT1 transporter, there was no difference in the intracellular pools
of GFP-CAT1 in normal (Figure 4-5 A) and LPI (Figure 4-5 B) fibroblasts. In both cell
lines, the GFP-CAT1 fusion protein was detected on the plasma membrane, throughout
the cytoplasm in small vesicles, and highly concentrated in the perinuclear region
(probably representing Golgi).
Examination of organelle integrity in normal and LPI fibroblasts. As discussed,
no significant differences in the distributions of any of the amino acid transporters were
identified in the normal and LPI fibroblasts. However, given the differences in
morphology observed between normal and LPI fibroblasts and the presence of the LPI-
specific vacuoles, it is clear that the disorder is associated with a basic cellular defect that
can be visualized as a structural deformity in one or more of the organelles of the LPI
cells. Therefore, staining patterns generated by organelle-specific antibodies, in normal
and LPI cells, were compared in order to gain important information regarding the
integrity of the organelles in the LPI fibroblasts. A variety of organelle-specific
antibodies were used as markers for the identification of compartments involved in the
biosynthetic, endocytic, and degradative pathways. The antibodies that were chosen are
prototypical markers for membrane protein trafficking compartments and have been


3
Kilberg, 1996). and are related to two members of the "ASCT family" that mediate Na -
dependent transport of selected zwitterionic amino acids (Arriza et al., 1993; Shafqat et
al., 1993; Liao and Lane, 1995). The NaVCl'-dependent proline transporter, as well as all
four glycine transporting members of the "GLYT family," are components of a large
superfamily of neurotransmitter transporters and are primarily expressed in the central
nervous system (reviewed by Malandro and Kilberg, 1996). Two cDNAs, designated
NBAT and 4F2hc, express proteins with only one to four transmembrane spanning
domains and comprise a unique family responsible for cationic and zwitterionic amino
acid transport (reviewed by Palacin, 1994; Malandro and Kilberg, 1996). However, it has
not yet been determined whether these proteins are actual transporters themselves, or
rather function as regulators or accessory subunits of a transporter complex.
Cationic Amino Acid Transport Systems
System y+, the primary mechanism for the transport of cationic amino acids, was
first described by White and Christensen in the early 1980s (White and Christensen,
1982; White et al.. 1982). This transport activity, first characterized in the Ehrlich cell,
was shown to be Na-independent, pH-insensitive, and stereoselective for L-amino acids
(reviewed by Kilberg et al., 1993). The tissue distribution was widespread, yet not
ubiquitous, and amino acids accumulated against a concentration gradient in response to
membrane potential. The uptake of amino acids was also subject to trans-stimulation, the
property of increased amino acid transport when substrate concentrations are elevated on
the opposite side of the membrane. Lastly, zwitterionic amino acids in the presence of
Na ions could competitively inhibit the transport of cationic amino acids by System y^


62
Q>
&
100,000g pellet
0)
o
CONFLUENCE
o
CO
50 75 100%
0

203
118
Figure 3-6. Immunoblot analysis of EAAT1 in the nuclear and intracellular membrane
fractions from human fibroblasts. A 30 ug aliquot of the 300 x g nuclear fraction (lane
1) and the 100,000 x g total intracellular membrane fraction (lanes 2-4) was subjected to
SDS-PAGE as described in the Methods Section of Chapter 3. Immunoblot analysis was
performed with a 1:1,000 dilution of EAAT1-R antibody and was detected with a
1:10,000 dilution of goat anti-rabbit IgG conjugated to horseradish peroxidase (HRP).
The 50, 75, and 100% labels indicates the degree of cell confluence at the time of cell
lysis and subfractionation. The blot shown is representative of three independent
experiments.


CHAPTER 3
DISTRIBUTION OF THE GLUTAMATE TRANSPORTERS
Introduction
Five plasma membrane proteins have been cloned that belong to the family of
transporters responsible for the Na'-dependent uptake of glutamate and aspartate in a
variety of tissues (see Chapter 1 for overview). These transporters correspond to the
high-affinity. Na~-dependent System XAG activity previously described in human
fibroblasts (Dall'Asta et al., 1983) and cultured liver cells (Makowske and Christensen.
1982). The glutamate transporters share approximately 50% amino acid identity, consist
of 6 or more transmembrane spanning domains, and contain two experimentally
identified N-linked glycosylation sites on the second extracellular loop (Conradt et al.,
1995). Glutamate uptake by these transport proteins is electrogenic. and coupled to the
co-transport of three Na~ ions and one H\ as well as the counter-transport of one Kr ion
(Zerangue and Kavanaugh. 1996).
Although similar in structure and substrate specificity, the glutamate/aspartate
transporters are differentially expressed and regulated. In the brain. EAAT1 (GLAST1)
and EAAT2 (GLT1) are localized to astroglia, whereas EAAT3 (EAAC1) is specific to
neurons. EAAT4 also has been localized to neurons and EAAT5 is specific to retinal
tissue. The family members also differ with respect to their transport properties. When
ooyctes were injected with EAAT5 cRNA, radiolabeled glutamate uptake was increased
28


126
pulmonary artery endothelial cells (PAEC), System y+ has been extensively characterized
(Zharikov and Block, 1997), and is responsible for 60-80% of total carrier-mediated
arginine uptake (Greene et al., 1993; McDonald et ah, 1997). In 1991, two laboratories
independently documented that the native biologic function of the previously cloned
murine ecotropic retroviral receptor was System y+ transport activity (Kim et ah, 1991;
Wang et ah, 1991). The mRNA and corresponding protein, termed CAT1, are expressed
in a wide variety of cells, with the notable exception of liver (Kim et ah, 1991; Wang et
ah, 1991; Wu et ah, 1994; Kakuda et ah, 1993). Using a CAT1 antibody generated in our
laboratory, it was shown that the arginine transporter is concentrated in specific regions
of the plasma membrane in a number of cell types, including PAEC (Woodard et ah,
1994; McDonald et ah, 1997).
Considerable information has been published regarding the presence of plasma
membrane micro-domains referred to as caveolae (Parton, 1996; Simionescu and
Simionescu, 1987; Schnitzer et ah, 1994; Anderson, 1993; Lisanti et ah, 1994). These
specialized membrane regions contain one or more of a family of structural proteins
called caveolins, as well as numerous signaling proteins, such as G-protein coupled
receptor systems, and a high cholesterol content. Of the three known isoforms, caveolin-
1 and caveolin-2 are most abundantly expressed in fibroblasts, endothelial cells, and
adipocytes, whereas caveolin-3 exhibits muscle-specific expression (reviewed by
Okamoto et ah, 1998). In addition to concentrating specific proteins within a specialized
region of the plasma membrane, caveolin may play a role in regulating the activation of
associated proteins. Caveolin homo- and heterodimers (between caveolin-1 and -2) form
a membrane-embedded hairpin structure resulting in cytosolic N- and C-termini (Scherer


129
rapid palmitoylation, plasma membrane association, and finally, partitioning to detergent-
insoluble membrane sub-domains, possibly caveolae.
The hypothesis of this project is that, in PAEC, the CAT1 transporter-containing
clusters, mentioned above, represent plasma membrane caveolae, and co-localization of
CAT 1-mediated arginine transport and eNOS provides an efficient mechanism for
delivery of substrate for NO synthesis, perhaps even in a direct manner. Additionally, the
trafficking of CAT1 to the caveolae may provide an alternative regulatory mechanism for
the production of NO. The following experiments were performed to test for the co
localization of CAT1 and eNOS within PAEC caveolae, as well as to determine if a
targeting signal for caveolar localization exists in the CAT1 protein sequence. My
working hypothesis is that the CAT1 transporter is co-localized with eNOS in caveolae
and thus, directs arginine to the enzyme for a more effective synthesis of NO.
Methods
Immunofluorescence assay. The immunofluorescence assays in this chapter were
performed in triplicate according to the methodology in Chapter 2. The antibodies used
in this chapter are described in Table 5-1. The secondary antibodies used were goat anti
rabbit or goat anti-mouse IgG conjugated to either FITC or TEXAS RED. The secondary
antibodies were used at a dilution of 1:200, and all immunofluorescence assays in this
chapter were analyzed by deconvolution microscopy.


Figure 4-1. Morphology of normal and LPI fibroblasts by light microscopy. Normal (A)
and LPI (B) fibroblasts were grown under normal culture conditions according to the
protocol in the Methods Chapter and visualized in the 75-mm culture tray using a Nikon
Axiophot epifluorescence inverted microscope. The digitized image was captured using a
Spot CCD camera with a resolution of 1315 x 1033 pixels (Diagnostics Instruments, Inc.,
Sterling Heights, MI). Fluorescence from three independent experiments was analyzed
and shown to be reproducible.


97


125
Palmer and Moneada, 1989; Pollock et al., 1991). It has been reported, by several
laboratories, that intracellular arginine concentrations range from 100 to 800 pM in
cultured endothelial cells (Block et al., 1995; Baydoun et al., 1990; Hecker et al., 1990;
Mitchell et ah, 1990; Gold et ah, 1989). Consequently, eNOS should be saturated in
these cells. Therefore, increasing the extracellular arginine should not increase NO
production any further. However, a number of in vitro and in vivo studies indicate that
NO production by vascular endothelial cells under physiological conditions can be
increased by extracellular arginine (Aisaka et ah, 1989; Cooke et ah, 1991; Taylor and
Poston, 1994; Rossitch et ah, 1991; Eddahibi et ah, 1992). Furthermore, a recent report
by Arnal et ah demonstrates that the intracellular concentration of arginine in endothelial
cells can be varied over 100-fold without changing NO production (Arnal et ah, 1995).
This observation, i.e., that extracellular arginine administration seems to drive NO
production even when intracellular levels of arginine are available in excess, has been
termed the arginine paradox and cannot be explained based on the available data (Kurz
and Harrison, 1997). One paradigm that would explain this observation is that in
endothelial cells the intracellular arginine is in one or more pools that are poorly, if at all,
accessible to eNOS, whereas extracellular arginine transported into the cell is
preferentially delivered to eNOS. Under this paradigm, a plasma membrane arginine
transporter might be in close spatial alignment with, or directly linked to, the plasma
membrane bound eNOS protein.
Arginine transport is mediated by several independent transport activities in
mammalian cells (White, 1985; Malandro and Kilberg, 1996; Closs, 1996). In porcine


Figure 3-5. Intracellular staining of Hela cells with EAAT1-R and EAAT1-S antibodies
Using the methods described in the Methods Chapter, Hela cells were fixed with -20C
MeOH and subjected to immunohistochemistry with antibodies specific for EAAT1-R
(A) and EAAT1-S (B). The EAAT1-R and EAAT1-S antibodies were detected by an
FITC-labeled goat anti-rabbit IgG secondary antibody. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
Panel A shows the nuclear staining of multiple cells, whereas panel B is one cell
(outlined) with intracellular vesicle staining. The data shown represent analysis of 0.2
pm sections through the cells.


iSi


135
Golgi (Sessa et al., 1995). Although there appears to be some co-localization of CAT1
and eNOS in the Golgi region, it is uncertain as to whether an interaction is already
present within the Golgi, or if the two proteins are just in close proximity as they advance
along the biosynthetic or recycling pathways. Further research is needed to clarify the
site of formation and the potential intracellular localization of the complex.
Paraformaldehyde-fixed PAEC were also dual stained with eNOS and a 1:50 dilution of
GLUT1 (see Table 4-1) or a 1:100 dilution of (3-integrin (see Table 4-1) to determine if
membrane proteins other than CAT1 co-localized with eNOS. Results from three
separate experiments demonstrated no significant overlapping fluorescence in either case
(data not shown).
Transfection of CAT1-GFP in PAEC. To confirm the results obtained from
immunofluorescence studies to detect endogenous CAT1, as well as to develop an
expression system to be used in future experiments, a GFP(C3)-CAT1 fusion protein was
constructed with the GFP tag at the N-terminus of the CAT1 sequence (see Methods
Chapter for details). PAEC were transfected with GFP(C3)-CAT1 or GFP(C3) vector
alone according to the Lipofectamine protocol (see Methods Chapter). Following 24 to
48 h of expression, PAEC were fixed with -20C MeOH and visualized by deconvolution
microscopy. In the absence of a targeting signal, the GFP(C3) shows diffuse
fluorescence throughout the cytoplasm and nucleus (data not shown). The GFP(C3)-
CAT1, on the other hand, is targeted specifically to the plasma membrane, and apparently
vesicle compartments involved in biosynthesis and intracellular trafficking as indicated
by a strong punctate pattern throughout the cytoplasm (Figure 5-5 A). Although the over-


[
SUBCELLULAR LOCALIZATION AND TRAFFICKING
OF AMINO ACID TRANSPORTERS
By
KELLY KRISTIN MCDONALD
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1998

ACKNOWLEDGMENTS
I would like to thank the members of my supervisory committee: Drs. Brian Cain,
William Dunn, Susan Frost, Mike Kilberg, and Peter McGuire. I wish to extend special
thanks to my mentor, Mike Kilberg, as well as Dr. Edward Block for his contribution to
the work discussed in Chapter 5. 1 would also like to acknowledge Dr. Stephen Wang for
his valuable instruction on the deconvolution microscope, David Parks, in the Center for
Structure Biology Computer Core, and Stephen Nowicki for valuable support and
friendship. Lastly, I would like to thank my parents, Drs. Maurice and Patricia
McDonald for their guidance, encouragement, and love.
li

TABLE OF CONTENTS
ACKNOWLEDGMENTS ii
LIST OF TABLES v
LIST OF FIGURES vi
ABSTRACT ix
CHAPTER 1 INTRODUCTION 1
Overview of Mammalian Amino Acid Transport 1
Trafficking of Membrane Proteins 11
Cell Biological Techniques for Studying Protein Trafficking 15
CHAPTER 2 MATERIALS AND METHODS 17
Materials 17
Methods 18
CHAPTER 3 DISTRIBUTION OF THE GLUTAMATE TRANSPORTERS 28
Introduction 28
Methods 36
Results 39
Discussion 48
CHAPTER 4 LYSINURIC PROTEIN INTOLERANCE 69
Introduction 69
Results 76
Discussion 89
CHAPTER 5 CAVEOLAR COMPLEX BETWEEN THE CATIONIC AMINO
ACID TRANSPORTER 1 AND ENDOTHELIAL NITRIC
OXIDE SYNTHASE 124
Introduction 124
Methods 129
Results 132
iii

Discussion 143
CHAPTER 6 CONCLUSIONS AND FUTURE DIRECTIONS 169
LITERATURE CITED 173
BIOGRAPHICAL SKETCH 185
IV

LIST OF TABLES
Table page
1-1. cDNA Clones of Amino Acid Transporters in CAT and EAAT Families 2
3-1. Antibodies against Glutamate Transporters 39
4-1. Antibodies for Immunofluorescence Studies 78
5-1. Antibodies for Immunofluorescence Studies 130
5-2. Immunodepletion of CAT 1-mediated Arginine Transport Activity
by anti-eNOS Antibody 139
v

LIST OF FIGURES
Figure page
3-1. Extracellular staining of human fibroblasts with EAAT3 antibody 53
3-2. Intracellular staining of human fibroblasts with EAAT3 antibody
and co-localization with organelle-specific antibodies 55
3-3. Nuclear staining of human fibroblasts with EAAT1 -R antibody and
co-localization with nucleus-specific antibodies 57
3-4. Intracellular staining of human fibroblasts with EAAT1 -S and
EAAT1-C antibodies 59
3-5. Intracellular staining of Hela cells with EAAT1-R and EAAT1-S
antibodies 61
3-6. Immunoblot analysis of EAAT1 in the nuclear and intracellular
membrane fractions from human fibroblasts 62
3-7. Expression of GFP and GFP-EAAT1 fusion proteins in human
fibroblasts 64
3-8. EAAT1 immunofluorescent staining of human fibroblasts
transfected with EAAT1-GFP(N3) 66
3-9. EAAT1 immunofluorescent staining of PAEC transfected with
EAAT1-GFP(N3) 68
4-1. Morphology of normal and LPI fibroblasts by light microscopy 97
4-2. Morphology of normal and LPI fibroblasts by electron microscopy 99
4-3. Intracellular staining of normal and LPI human fibroblasts with the
CAT1 antibody 101
4-4. Expression of GFP and the GFP(C3)-CAT1 fusion protein in
normal human fibroblasts 103
4-5. Expression of the GFP(C3)-CAT1 fusion protein in normal and
LPI human fibroblasts 105
vi

4-6. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the nuclear membrane and nucleolus 107
4-7. Intracellular staining of normal and LPI fibroblasts with antibodies
against plasma membrane and cytoskeletal proteins 109
4-8. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the endoplasmic reticulum and Golgi Complex Ill
4-9. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the endocytic and recycling pathways 113
4-10. Intracellular staining of normal and LPI fibroblasts with an
antibody against a lysosomal enzyme 115
4-11. Intracellular staining of normal and LPI fibroblasts with an
antibody against a lysosomal membrane protein 117
4-12. Visualization of acidic compartments of normal and LPI
fibroblasts with acridine orange 119
4-13. Lysosomal detection in normal and LPI fibroblasts following
chloroquine treatment 121
4-14. Lysosomal staining of normal and LPI cells expressing the
GFP(C3)-CAT1 fusion protein 123
5-1. Surface labeling of PAEC with the CAT1 transporter antibody 149
5-2. Co-localization of CAT1 and caveolin on PAEC 151
5-3. Co-localization of CAT1 and eNOS on PAEC 153
5-4. Detection of CAT1 and eNOS in the Golgi of PAEC 155
5-5. Expression of the GFP(C3)-CAT1 fusion protein in PAEC 157
5-6. Immunofluorescent staining of PAEC transfected with GFP(C3)-
CAT1 fusion protein 159
5-7. Distruption of CATl/eNOS co-localization in PAEC treated with
nocodazole 161
5-8. CAT1 immunoprecipitation of eNOS from solubilized PAEC
plasma membrane vesicles 162
vii

5-9. Immunofluorescent staining of PAEC transfected with the
CATMUT1 palmitoylation mutant 164
5-10. Immunofluorescent staining of PAEC transfected with the
CATMUT3 palmitoylation mutant 166
5-11. eNOS staining of PAEC treated with varying concentrations of
extracellular L-arginine 168
vm

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SUBCELLULAR LOCALIZATION AND TRAFFICKING
OF AMINO ACID TRANSPORTERS
By
Kelly Kristin McDonald
August, 1998
Chairman: Dr. Michael S. Kilberg
Major Department: Biochemistry and Molecular Biology
Mammalian amino acid transporters have been well-characterized with regard to
substrate specificity, kinetic parameters, and metabolic regulation. However, little
information is known concerning the cell biology or "life cycle" of plasma membrane
amino acid transporters. In this study, molecular and cell biological techniques, including
immunohistochemistry, transporter mutants and expression of green fluorescent protein
fusions, and deconvolution microscopy, were used to examine the cellular localization
and trafficking of specific amino acid transporters under normal and diseased conditions.
The availability of cDNAs and antibodies for specific members of the EAAT anionic and
the CAT cationic amino acid transporter families have provided an avenue for comparing
the subcellular distribution of amino acid transporters from the same, as well as different
gene families. Although similar in structure and function, the EAAT1 glutamate
transporter was detected primarily in the nuclear membrane in certain cell types, such as
IX

human fibroblasts, whereas the EAAT3 transporter was observed in intracellular vesicle
compartments and in concentrated clusters on the plasma membrane. Lysinuric Protein
Intolerance (LPI) provided a model system for investigating the subcellular organization
and trafficking pathways in a disease with visual morphological defects. A survey of
organelle integrity led to the discovery of an abnormal population of lysosomes in the
LPI fibroblasts. Future studies will investigate the contents of the LPI lysosomes, as well
as the amino acid transport across lysosome-enriched membrane preparations.
Immunostaining of porcine aortic endothelial cells (PAEC) revealed "patches" of the
cationic amino acid transporter CAT1 that co-localized with antibodies against caveolin
and endothelial nitric oxide synthase (eNOS). When incubated with solubilized PAEC
plasma membrane proteins, an eNOS-specific antibody immunoprecipitated CAT1-
specific arginine transport activity. These results document the existence of a caveolar
complex between CAT1 and eNOS in PAEC that provides a potential mechanism for the
efficient delivery of the arginine substrate to eNOS for nitric oxide (NO) production. The
individual projects presented in this thesis share a common goal in documenting the
cellular localization, and when possible, understanding the functional consequences, of
the transporter protein distribution.
x

CHAPTER 1
INTRODUCTION
Overview of Mammalian Amino Acid Transport
Over the past three decades, a variety of mammalian amino acid transport systems
have been characterized extensively for cell and substrate specificity, kinetic parameters,
and metabolic regulation. However, the original studies began with the pioneering work
of Van Slyke and Meyer in 1913, when they demonstrated that tissues accumulated
amino acids against a concentration gradient (Van Slyke and Meyer, 1913). In the early
1960s, the Christensen laboratory began to define specific transport systems that mediate
the flux of amino acids across the membrane bilayer based on the conformation, size, and
chemical properties of the amino acid side chain (Oxender and Christensen, 1963).
Rigorous investigation revealed that each system recognizes more than one amino acid,
distinct systems exhibit some degree of overlapping substrate reactivity, and various
amino acids are transported by more than one system. These observations were further
complicated by the discovery that distinct cell and tissue types express a different
combination of systems that work in concert to provide specific nutritional requirements.
The identification of individual proteins responsible for amino acid transport activity has
been complicated by the relatively low abundance of these proteins, as well as the
technical difficulty involved in isolating and purifying integral membrane proteins.
However, the advances in molecular biology over the last ten years have resulted in the
1

2
cloning and expression of more than 20 cDNAs encoding amino acid transporters. Based
on sequence homology, a number of the cloned transporters have been classified
according to gene families. Those related to the studies presented in this thesis are
summarized in Table 1-1.
Table 1-1
cDNA Clones of Amino Acid Transporters in CAT and EAAT Families
Clone
Alternate
names
Deduced amino
acid length
Substrate
specificity
Ions coupled
CAT1

622-624
cationic

CAT2
CAT2a
658
cationic

CAT2a
CAT2b
659
cationic

CAT3

619
cationic

EAAT1
GLAST1
GluT
543
D,L-aspartate
L-glutamate
Na jn, K. out
H+in
EAAT2
GLT1, GLTR
GLAST2
573
D,L-aspartate
L-glutamate
Na+,, K+oul
H|n
EAAT3
EAAC1
523-525
D,L-aspartate
L-glutamate
Na+in, K+out
EAAT4
564
D,L-aspartate
L-glutamate
Na+in, K+out
H+in
EAAT5
560
D,L-aspartate
L-glutamate
Na+jn, K+oul
H,n
The "CAT family" is composed of four known transporters that mediate the Na+-
independent transport of the cationic amino acids arginine, lysine, ornithine, and histidine
when positively charged (reviewed by MacLeod et al., 1994; Malandro and Kilberg,
1996). Although the members of this family share a common substrate specificity and
significant amino acid similarity, they differ in tissue distribution, affinity for substrate,
and most likely the functional role they play in metabolism. The anionic transporter
family currently contains five members that are responsible for the Na+-dependent
transport of the anionic amino acids glutamate and aspartate (reviewed by Malandro and

3
Kilberg, 1996). and are related to two members of the "ASCT family" that mediate Na -
dependent transport of selected zwitterionic amino acids (Arriza et al., 1993; Shafqat et
al., 1993; Liao and Lane, 1995). The NaVCl'-dependent proline transporter, as well as all
four glycine transporting members of the "GLYT family," are components of a large
superfamily of neurotransmitter transporters and are primarily expressed in the central
nervous system (reviewed by Malandro and Kilberg, 1996). Two cDNAs, designated
NBAT and 4F2hc, express proteins with only one to four transmembrane spanning
domains and comprise a unique family responsible for cationic and zwitterionic amino
acid transport (reviewed by Palacin, 1994; Malandro and Kilberg, 1996). However, it has
not yet been determined whether these proteins are actual transporters themselves, or
rather function as regulators or accessory subunits of a transporter complex.
Cationic Amino Acid Transport Systems
System y+, the primary mechanism for the transport of cationic amino acids, was
first described by White and Christensen in the early 1980s (White and Christensen,
1982; White et al.. 1982). This transport activity, first characterized in the Ehrlich cell,
was shown to be Na-independent, pH-insensitive, and stereoselective for L-amino acids
(reviewed by Kilberg et al., 1993). The tissue distribution was widespread, yet not
ubiquitous, and amino acids accumulated against a concentration gradient in response to
membrane potential. The uptake of amino acids was also subject to trans-stimulation, the
property of increased amino acid transport when substrate concentrations are elevated on
the opposite side of the membrane. Lastly, zwitterionic amino acids in the presence of
Na ions could competitively inhibit the transport of cationic amino acids by System y^

4
(reviewed by Kilberg et al., 1993). Presently, four distinct proteins that specifically
mediate plasma membrane cationic amino acid transport have been cloned, and
demonstrate similar, yet not identical, properties to those of System y+.
I have chosen to study the primary cationic amino acid transporter, designated
CAT1, as well as several members of the glutamate family (described below). In 1989,
Cunningham and coworkers identified a cDNA clone that encoded the ecotropic murine
leukemia virus receptor (Albritton et al., 1989). It was later determined that the receptor
also functioned as a Na'-independent, high-affinity transporter for arginine, lysine,
ornithine, and histidine when positively charged (Kim et al., 1991; Wang et al., 1991).
The corresponding human cDNA was cloned by Meruelo and coworkers (Yoshimoto et
al., 1991) and is 88% homologous to the murine transporter, but lacks homology in the
region of viral binding. This is consistent with the observation that the ecotropic murine
leukemia retrovirus is unable to infect human cells (Albritton et al., 1993). The human
CAT1 sequence revealed a 629-amino acid protein with 12-14 predicted transmembrane
spanning domains. The CAT1 transporter mRNA is not expressed in the liver, but
otherwise appears to be ubiquitous (for review see Malandro and Kilberg, 1996).
Immunohistochemistry from our laboratory, using an antibody against the third
extracellular loop of the murine cDNA sequence, confirmed the lack of expression in rat
liver (Woodard et al., 1994).
MacLeod and coworkers cloned a second member of the CAT family (MacLeod
et al., 1990) from murine T cell lymphocytes (Kekuda et al., 1993). This protein,
originally called the Tea gene, was discovered during a search for genes involved in T
cell activation, however, its function as a transporter was suggested by its extensive

5
sequence homology to CAT1. The Tea gene, later renamed murine CAT2, is 61%
identical to CAT1 at the amino acid level, has 12 predicted membrane spanning domains,
and mediates Na'-independent high-affinity cationic amino acid transport in activated T
lymphocytes (for review see Malandro and Kilberg, 1996). Chimeric constructs of the
mouse and human CAT1 sequences were used to identify the region of viral binding in
the third extracellular loop of the murine CAT1 protein (Albritton et al., 1993). The
CAT2 sequence, like the human CAT1 sequence, is divergent in this region and does not
function as a binding site for the retrovirus. Nitric oxide (NO) has been implicated in T
cell signaling by autocrine/paracrine pathways, and it has been proposed that the
expression of the CAT2 transporter during T cell activation may be related to the need for
arginine in the production of NO (MacLeod et al., 1994). Using the murine CAT2
sequence as a probe for screening a mouse liver cDNA library, Cunningham and
coworkers identified a third liver-specific member of the CAT family, murine CAT2a.
This protein is encoded by the same gene as CAT2, but as a result of differential splicing,
CAT2a has an additional stretch of 41 amino acid residues (358-398) between the eighth
and ninth membrane spanning domains. Despite the similarity in sequence, CAT2a
exhibits a 10-fold lower affinity for arginine than either CAT1 or CAT2 (Closs et al.,
1993). This kinetic difference suggests that the 41 amino acid region may be involved in
binding the amino acid during its translocation across the membrane.
Screening a rat brain cDNA library with probes designed from the murine CAT1
sequence isolated the most recent addition to the CAT family, CAT3 (Hosokawa et al.,
1997). The protein encoded by the rat CAT3 is comprised of 619 amino acids and shares
53-58% identity with the CAT family members previously described. The CAT3 protein

6
mediates the high-affinity, Na*-independent transport of cationic amino acids and shares
the greatest homology with the CAT1 family member. In the same year, the mouse
CAT3 cDNA was identified by Ito and Groudine during an attempt to isolate germ-layer
specific transcripts from mouse embryos (Ito and Groudine, 1997). The mouse CATS
was localized to the brain by in situ hybridization studies, and exhibited the same
structural and transport characteristics as the rat homolog. The highly conserved tissue-
specificity between the rat and mouse proteins suggests an important role for CAT3 in the
brain. Detection of mRNA in brain capillaries suggests a role for CAT3 in the transport
of cationic amino acids by endothelial cells at the blood-brain barrier (Hosokawa et al.,
1997). Just as CAT2 may provide the inducible nitric oxide synthase (iNOS) with
substrate for NO production in T cells, it is feasible that CAT3 provides the neuronal
nitric oxide synthase (nNOS) isoform with the arginine required for NO production in the
nervous system.
The CAT family, comprising four functionally similar yet distinct proteins, is
only one of two families responsible for the transport of cationic amino acids. The other
family has been identified based on mRNA expression in Xenopus oocytes and currently
includes two proteins, NBAT and 4F2hc. Expression of either in oocytes induces Na"-
independent transport of cationic amino acids, but NBAT mediates the uptake of Na+-
independent zwitterionic amino acids whereas 4F2hc-catalyzed uptake of these substrates
is Na-dependent (Bertrn et al., 1992; Wells et al., 1992). These specificities correspond
to two known transport activities, called System b+ (NBAT) and System y+L (4F2hc),
respectively. The hydropathy plots of both transporters predict either one or four
membrane spanning domains. This is an interesting feature for two proteins that are

7
believed to function in transport, because most of the cloned transporters are thought to
span the membrane 12-14 times. That NBAT and 4F2hc have four or less trans
membrane domains has led to the hypothesis that they may serve as modulators, or
comprise only one subunit of a multimeric transporter complex (for review see Palacin,
1994). Co-precipitation and cross-linking experiments have recently provided strong
evidence that NBAT, a 90 kDa protein, is associated with a 50 kDa protein (Wang and
Tate, 1995). The NBAT protein is expressed in the microvilli of proximal tubules of the
kidney and the mucosa of the small intestine (Kanai et al., 1992). Expression of NBAT
in Xenopus oocytes has been shown to induce the high-affinity uptake of cystine. This
finding provided key evidence for NBAT's involvement in cystinuria, an autosomal
recessive disorder characterized by the hyperexcretion of cystine and cationic amino acids
into the urine (Segal and Thier, 1989; reviewed by Palacin, 1994). At least six distinct
missense mutations in NBAT have been documented in different patients with the
disease, but all result in defective transport of cystine through the epithelial cells of the
renal tubule and intestinal tract (Calonge et al., 1994).
Anionic Amino Acid Transport Systems
The anionic transporter family currently includes at least five members that
mediate glutamate/aspartate transport. These Na+-dependent transporters share a similar
structure with six predicted trans-membrane spanning domains in the N-terminal portions
and a large hydrophobic region at the C-termini that may represent additional trans
membrane domains. Each transporter has at least two putative glycosylation sites and
shares from 40-68% amino acid sequence identity with the other members of the family.

8
Glutamate transport studies in salamander retina glial cells demonstrated that glutamate
uptake is electrogenic and coupled to the co-transport of three Na* ions and a FT, as well
as the counter-transport of one K ion (Zerangue and Kavanaugh, 1996). In the central
nervous system, the stoichiometry of the glutamate and ion transport must be tightly
regulated. Ischemic conditions following a stroke may lead to the breakdown of
electrochemical gradients as a result of lower ATP levels and reduced functioning of
Na\K ATPase proteins (Szatkowski and Attwell, 1994). If the ion gradients are
disrupted, then it is believed that the glutamate transporters can function in reverse,
resulting in the release of glutamate into the synaptic cleft, and subsequent neurotoxicity
and neuronal death (Kanai et al., 1995).
In an attempt to isolate galactosyltransferase from rat brain, Storck and coworkers
co-purified the first glutamate transporter, designated GLAST1 (Storck et al., 1992). The
isolated protein showed homology to the previously cloned bacterial glutamate and
monocarboxylate transporters, and its function as a glutamate transporter was confirmed
by expression in Xenopus oocytes followed by radiolabeled amino acid uptake (Klockner
et al., 1993). Pines and coworkers cloned the second glutamate transporter, GLT1, by
screening a rat cDNA expression library with antibodies generated against the partially
purified protein (Pines et al., 1992). Northern analysis and in situ hybridization detects
GLAST1 and GLT1 mRNA expression primarily in glial cells of the central nervous
system (Storck et al., 1992; Otori et al., 1994), where they play an important role in
clearing toxic levels of glutamate from the synaptic clefts. Although the mechanism is
unclear, recent data from Rothstein and coworkers indicate that abnormal GLT1 mRNA
species may be responsible for the decreased glutamate transport detected in patients with

9
Amyotrophic Lateral Sclerosis (ALS), but the defective glutamate transport is not thought
to be the principle cause of the disease (Lin et al., 1998). The third member of the
growing glutamate transporter family, EAAC1, was cloned by oocyte expression cloning
using fractionated mRNA from rabbit intestine (Kanai and Hediger, 1992). Although
similar to GLAST1 and GLT1 in structure and transport properties, EAAC1 expression is
neuron-specific in the brain. EAAC1 transcripts have also been detected in small
intestine, kidney, liver, heart, placenta, and skeletal muscle (reviewed by Malandro and
Kilberg, 1996). In 1994, the human homolog to GLAST1 was identified (Arriza et al.,
1994; Kawakami et al., 1994) and given the name EAAT1, for Excitatory Amino Acid
Transporter 1. Shortly after, the human GLT1 sequence was reported (Arriza et al., 1994;
Manfras et al., 1994) and designated EAAT2, and the human homolog to EAAC1 was
cloned (Arriza et al., 1994; Kanai et al., 1995) and called EAAT3. It has since become
acceptable to refer to the glutamate transporters by the EAAT nomenclature regardless of
the species.
EAAT4 was isolated using degenerative oligonucleotide primers corresponding to
conserved sequences within the other members of the glutamate family (Fairman et al.,
1995). Northern analysis identified EAAT4 mRNA in the cerebellum and placenta, and
transport studies in oocytes demonstrated a high-affinity glutamate uptake that was
associated with chloride conductance. The final member of this family, EAAT5, was
cloned by Arriza et al. by screening a human retinal cDNA library with a glutamate
transporter cDNA isolated from salamander retina (Arriza et al., 1997). Like EAAT4,
EAAT5 may play a role in ion conductance instead of, or in addition to, providing
neurotransmitter clearance at the synaptic cleft. Electrophysiological studies of both

10
EAAT4 and EAAT5 have shown a large chloride conductance in addition to transport
activity. In the case of EAAT5, this associated chloride conductance may participate in
visual processing (Arriza et al., 1997). A potential PSD-95-binding motif was identified
in the C-terminus of EAAT5 (Arriza et al.. 1997). PSD-95, a cytoskeleton-associated
synaptic protein, has been shown to bind to C-terminal sequences in both the N-methyl-
D-aspartate (NMDA) receptor and Shaker-type potassium channels (Cho et al., 1992).
ASCT Amino Acid Transport Systems
Two zwitterionic amino acid transporters, ASCT1 (Shafqat et al., 1993; Arriza et
al., 1993) and ASCT2 (Kekuda et al., 1996; Utsunomiya-Tate et al., 1996), are 56%
identical to one another at the amino acid level, and both are approximately 40% identical
to the five glutamate transporters. The "ASC" transporters are Na-dependent, and
although they exhibit a broad substrate specificity, they prefer amino acids with
hydroxyl- or sulfydryl-containing side chains (serine, cysteine, and threonine). Like
System ASC, the transporters exhibit the property of trans-stimulation even though this
activity is typically a characteristic of Na+-independent transporters. Northern blot
analysis revealed highest expression of ASCT1 in the brain, skeletal muscle, and pancreas
(Shafquat et al.. 1993; Arriza et al., 1993). ASCT2, also known as ATB, mRNA is
expressed in lung, skeletal muscle, kidney, large intestine, testes, and adipose tissue
(Kekuda et al., 1996; Utsunomiya et al., 1996). This transporter, only recently cloned by
homology to ASCT1, has been shown to exhibit a broader substrate specificity than
ASCT1 (Kekuda et al., 1996; Utsunomiya et al., 1996). In fact, one of the distinguishing

features between the two is the acceptance of glutamine by ASCT2, but not by ASCT1
(Utsunomiya et al., 1996).
ASCT1 exhibits a unique pH-dependent substrate specificity (Tamarappoo et al.,
1996), first described for System ASC by Christensen and colleagues (Vadgama and
Christensen, 1984). At neutral pH, ASCT1 preferentially transports zwitterionic amino
acids, whereas if the assay pH is lowered to 5.5, the transporter will accept both
zwitterionic and anionic amino acids. Transport assays using anionic amino acid analogs
to compete with the zwitterionic substrates at pH 5.5, indicate that the substrates may
share the same binding region (Tamarappoo et al., 1996). One hypothesis to explain this
pH effect is that one or more of the eight histidine residues in ASCT1 accept(s) a positive
charge when the pH is lowered below 6.0 (the pKa for the histidine side chain). The
positively charged histidine(s) may serve as a binding site for negatively charged amino
acids, such as glutamate, aspartate, and cysteate. ASCT2 also has a low affinity for
glutamate at neutral pH. and the affinity for the anionic amino acid increases as the assay
pH is lowered (Utsunomiya et al., 1996; Kekuda et al., 1996).
Trafficking of Membrane Proteins
The cloning and expression of a number of the amino acid transporters have led to
the generation of sequence-specific antibodies from corresponding peptides and fusion
proteins. The availability of antibodies and the development of high-resolution
microscopes have allowed investigators to initiate studies on the cell biology of amino
acid transporters, an area of the field in which little is known. The first mammalian
amino acid transporter antibody was produced in 1992. In recent years, many of the

12
individual steps that contribute to the "life cycle" of plasma membrane proteins have been
documented. Although little information is available, it is likely that many similarities
exist between the "life cycle" of amino acid transporters and other plasma membrane
proteins. From biogenesis to degradation, these proteins are transported through a
complex system of membrane compartments and organelles by specific vesicles that bud
from a donor membrane and fuse with a target membrane (Ivessa et al., 1995). A
combination of coat proteins and several classes of monomeric GTPases (i.e., Rabs) are
believed to regulate certain steps in vesicle trafficking. The SNARE hypothesis describes
the mechanism by which transport vesicles target membranes (Alberts et ah, 1994). v-
SNAREs are proteins on the vesicle membranes and t-SNAREs reside on the target
membranes. v-SNAREs and t-SNAREs are suspected to function as structural proteins
that interact at the point of vesicle docking. Over 30 different Rab proteins have been
identified and each is believed to play some role to ensure the specificity of individual
vesicle docking/fusion events of the membrane trafficking pathways (Nuoffer and Balch,
1994).
Three primary trafficking pathways have been described for various membrane
proteins: the biosynthetic-secretory pathway, the endocytic-exocytic pathway, and the
degradative pathway. In the biosynthetic pathway, integral membrane proteins and
secretory proteins are co-translationally inserted into the membrane or the lumen,
respectively, of the endoplasmic reticulum (ER) where early oligosaccharide modification
and proteolytic processing begin. Select proteins receive fatty acylation moieties either
during or following, the translation event (reviewed by Solski et al., 1995). Myristic and
palmitic acid modifications contribute to membrane binding and stability, and more

13
recently, have been implicated in targeting certain proteins to plasma membrane caveolae
(Song et al., 1996). Following translation, proteins are packaged into vesicles and
transported to the Golgi apparatus for the completion of glycosylation and folding events.
The developing proteins are shuttled from the cis, to the medial, to the trans-Golgi
compartment, ultimately arriving at the trans-Golgi network (TGN) where they are sorted
according to their final destinations (Alberts et al., 1994).
Membrane proteins also participate in the endocytic/exocytic pathway. Several
distinct forms of endocytosis have been described (reviewed by Watts and Marsh, 1992;
Alberts et al., 1994). During pinocytosis, small invaginated plasma membrane vesicles of
less than, or equal to 150 nm in diameter, constituatively carry fluids and solutes into the
cell. Phagocytosis, on the other hand, results in the regulated ingestion of large particles
via plasma membrane derived vesicles of greater than 250 nm in diameter, and is
generally the responsibility of specialized cells. It is assumed that there is an
internalization of many plasma membrane proteins during these two processes. Most
animal cells take up specific macromolecules by a process called receptor-mediated
endocytosis. During this event, receptor-ligand complexes participating in this cycle are
internalized by clatherin-coated pits on the plasma membrane and delivered to early
endosomes. The acidic environment of the endosme results in the dissociation of the
ligands, which advance via late endosomes to ultimate lysosomal degradation. Some
receptors are recycled directly back to the plasma membrane from the endosomal
compartment, whereas others recycle by way of an intermediate step in the TGN. The
transferrin receptor (TfR.) interacts with iron-bound transferrin at the plasma membrane
(reviewed by Hansen et al., 1993; Alberts et al., 1994). Following endocytosis, iron

14
molecules are released from the transferrin in response to the low pH of the endosme.
This allows the unbound transferrin to recycle to the plasma membrane with its receptor.
The transferrin molecule is then freed so that it may sequester more extracellular iron (De
Silva et al., 1996).
The final trafficking pathway that contributes to the life cycle of membrane
proteins involves the transport of proteins to the lysosomes for degradation (reviewed by
Komfeld and Mellman, 1989), or the ubiquitination and subsequent degradation by
proteosomes (reviewed by Hershko and Ciechanover, 1992). Materials from multiple
pathways are emptied into the lysosomes, where acid hydrolases function in the regulated
digestion of macromolecules such as membrane proteins. The mannose-6-phosphate
receptor (M6PR) binds to the phosphorylated mannose modification on lysosomal-
destined digestive enzymes and shuttles these acid hydrolases from the TGN to the late
endosomes. The acid hydrolases eventually end up in lysosomes and the M6PR recycles
to the TGN. Therefore, the M6PR is an excellent marker for the late endosomal and TGN
compartments. A second trafficking pathway that leads to degradation evolves from
endocytosis. In a process that is poorly understood, materials destined for degradation
are transferred from early endosomes to late endosomes to lysosomes (Alberts et al.,
1994). Antibodies against mammalian amino acid transporters have become available
only in the last couple of years, so no information has been published concerning the
molecular mechanisms by which these transporters are degraded.

15
Cell Biological Techniques for Studying Protein Trafficking
Indirect immunofluorescence has provided a way of studying the subcellular
localization and trafficking of proteins using a combination of antibodies and trafficking
inhibitors. Indirect immunofluorescence is a sensitive method for detecting a protein of
interest because many molecules of the secondary antibody recognize each molecule of
primary antibody, which is raised against a specific peptide or protein. This results in an
amplification of the signal because the secondary antibody is covalently attached to a
fluorochrome that fluoresces when exposed to a specific wavelength of light. However,
problems may arise if an antibody cannot be raised against a desired protein, or if an
antibody produces a high level of background by cross-reacting with other cellular
proteins or artifacts. To avoid some problems commonly associated with antibodies, and
to investigate living cells, a new technique that utilizes the autofluorescence of the green
fluorescent protein (GFP) has replaced, or is being used in conjunction with, antibody
labeling techniques.
GFP. a 27 kDa protein native to the bioluminescent jellyfish, Aequorea victoria,
produces a bright green color when stimulated by blue or UV light (reviewed by Steams,
1995). GFP expression is species-independent and can be introduced into prokaryotic
and eukaryotic cells without the requirement of specific cofactors, substrates, or
additional gene products. The GFP protein is small and globular, and in most cases does
not interfere with the synthesis, trafficking, or activity of the fusion protein product.
Whereas antibody labeling often involves the use of a fixative, GFP constructs can be

16
viewed in fixed or living cells. Certain GFP variants have been optimized for use in
mammalian cells and are now commercially available. These variants that contain
double-amino-acid substitutions Phe-64 to Leu and Ser-65 to Thr result in a 35-fold
increase in fluorescence over wild type GFP (Cormack et al., 1996). In addition,
expression has been enhanced by the introduction of silent mutations in the coding
sequence that correspond to human codon-usage preferences. The GFP serves as a
genetic tag that can be conveniently added to the protein coding sequence of a cDNA.
In this study, molecular and cell biological techniques, including
immunohistochemistry, GFP expression, and deconvolution microscopy, have been used
to examine specific aspects of the life cycle of amino acid transporters. The individual
projects presented in this thesis share a common goal in documenting the cellular
localization, and when possible, understanding the functional consequences, of
transporter distribution under normal and diseased conditions.

CHAPTER 2
MATERIALS AND METHODS
This Chapter contains the general materials and methods used during the course of
this research project. Chapter 3 and Chapter 5 include additional Methods Sections that
are specific for the work described in those chapters.
Materials
Fibronectin, bovine serum albumin (BSA), Triton X-100, and Triton X-l 14,
kanamycin, nocodozole, dimethyl sulfoxide, and adenosine triphosphate (ATP) were
purchased from Sigma Chemical Company (St. Louis, MO). RPMI and MEM media,
goat serum (NGS). fetal bovine serum (FBS), lipofectamine, OptiMEM, EcoRI and
Hindlll restriction enzymes and buffers, T4 polynucleotide kinase and buffer, and all
PCR primers and mutagenesis oligonucleotides were purchased from Gibco BRL
(Gaithersburg. MD). Paraformaldehyde (PFA), glycine and methanol (MeOH) were
purchased from Fisher Scientific (Pittsburgh, PA). Fluoromount-G was purchased from
Southern Biotechnology Associates (Birmingham, AL). The MORPH mutagenesis kit
was obtained from 5 Prime 3 Prime (Boulder, CO), the pCR2.1 TA cloning kit was
obtained from InVitrogen (Carlsbad, CA), and the DNA gel purification kit and DNA
plasmid purification kits were obtained from Quiagen (Valencia, CA.). The
nitrocellulose membranes were purchased from Cuno, Inc. (Meridian, CT) and the
enhanced chemiluminesence reagents were purchased from Pierce (Rockford, IL). The
17

18
MCAT1 cDNA was a generous gift from Dr. James Cunningham at Brigham and
Women's Hospital (Boston. MA), and the rat EAAC1 cDNA was cloned from a rat
hippocampal librar)' (Velaz-Faircloth et al., 1996). All of the antibodies are described in
the chapters in which they were used.
Methods
Cell culture. Pulmonary artery endothelial cells (PAEC) were prepared by
collaborators in Dr. Edward Block's laboratory. PAEC were isolated by collagenase
treatment of the main pulmonary artery of 6-7-month-old pigs and were cultured for 3-7
passages as described by Block and coworkers (Block et al., 1989). One hundred-mm
dishes or 6-well trays were incubated with 7-10 pg/ml of fibronectin (Sigma Chemical
Co.), dissolved in RPMI medium, overnight at 37C under a humidified atmosphere of
5% CO:-95% air. Prior to plating cells, the fibronectin solution was aspirated, and
dishes/trays were allowed to dry for 30 min in a culture hood under a UV lamp. Once
plated, PAEC were maintained in RPMI + 4% or 10% fetal bovine serum (FBS). Hela,
HepG2 (human hepatoma), HEK 293 (human embryonic kidney), CHO and BNL.CL2
(mouse hepatocytes) cells were maintained in Eagle's minimal medium (MEM) + 4%
FBS. Cultured fibroblasts from Finnish LPI patients and sex-age-matched normal
controls were obtained from Dr. Olli Simell at the Central Hospital, University of Turku
(Turku, Finland). The fibroblasts were cultured in 75-mm flasks and maintained in
MEM. supplemented with 10% FBS. The cells were passaged after achieving
approximately 80% confluence, and were used for experiments until the sixth to eighth

19
passage. Incubation of all cell lines described was at 37C under a humidified atmosphere
of 5% C02-95% air.
Immunohistochemistrv. For immunofluorescence assays, cells were transferred to
22 x 22 mm sterilized Corning glass microscope cover slips, by placing the cover slips in
the wells of the Falcon six-well cluster trays, and plating the cells, which were then
allowed to reach 60 to 70% confluence. PAEC cells required pre-treatment of the cover
slips with 7-10 fig/ml flbronectin, as described above for culture dishes. Following three
5-min washes with phosphate buffered saline (PBS) to remove culture medium, cells
were fixed with a 4% paraformaldehyde solution for 20 min. To prepare the fixation
solution, 4% paraformaldehyde was added to 50% of the final volume of water and the
mixture was heated to 60-65C. Drop-wise addition of 10N sodium hydroxide (usually 1-
2 drops) was required to completely dissolve the paraformaldehyde. After the solution
had cooled, 30% of the final volume of water and 20% of the final volume of 5X PBS
were added, and the solution was adjusted to a pH of 7.5 to 8.0. After incubation with the
cells, the fixative was removed with three 5-min washes in PBS, and any residual
paraformaldehyde was blocked with 50 mM glycine (in PBS) for 30 min. This
incubation was followed by three additional 5-min PBS washes. Paraformaldehyde
fixation was used for immunofluorescence experiments designed to examine plasma
membrane labeling. If cell permeabilization was desired in combination with
paraformaldehyde fixation, 0.1 to 0.2% Triton X-100 was added to the wells for the last
30 minutes of blocking. Alternatively, a -20C methanol incubation for 5-min was used
to fix cells for the purpose of intracellular staining. Following either fixation procedure,

20
cells were incubated in a solution of PBS containing 20% normal goat serum (NGS) and
3% bovine serum albumin (BSA), for 1-2 h in order to block non-specific antibody
binding. For pre-immune or immune labeling, cover slips were removed from wells and
inverted onto 50 pi drops of pre-immune or primary antibody solution. The primary
antibodies were prepared in 20% NGS/PBS with 3% BSA (plus Triton-X when
appropriate), and for peptide competition assays, allowed to incubate with 50 pg/ml of
corresponding peptide overnight at 4C. All incubations were at room temperature
(unless otherwise indicated), and antibody reactions were performed on parafilm in a dark
humid box. The dark humid box was prepared by placing PBS-saturated gauze across the
bottom of a small Tupperware container. Following a 2 h pre-immune or primary
antibody incubation, cells were returned to the wells and washed three times with PBS.
The secondary antibody, prepared in 20% NGS/PBS with 3% BSA (plus 0.1-0.2%
Triton-X when appropriate), was applied in a similar manner to the primary, and allowed
to incubate for 1 h before unbound molecules were removed with three 5 min PBS
washes. Following the last wash, cover slips were mounted onto glass slides with a drop
of Fluoromount-G, allowed to dry, and the edges of the cover slip sealed with fingernail
polish.
The secondary antibodies used in the immunofluorescence assays were either goat
anti-mouse or goat anti-rabbit IgG conjugated to either Texas Red or FITC (fluorescein
isothiocyanate) fluorochromes (unless otherwise stated). Texas Red is excited when
exposed to a wavelength of 593 nm and emits a red fluorescence at a wavelength of 612
nm. FITC is excited at 494 nm and emits a green fluorescence at 517 nm. Because of the

21
different fluorescent properties, these two secondary antibodies can be used in double
labeling experiments to show co-localization of two different proteins. However, the
proteins of interest must be detected using primary antibodies generated in two different
species. For example, double-labeling can be performed by incubating cells with a
primary antibody raised in mouse, and detected with a secondary goat anti-mouse Texas
Red antibody, and with another primary antibody raised in rabbit, detected using a goat
anti-rabbit FITC antibody.
Fluorescence light microscopy. Slides were initially viewed using a Nikon
Axiophot epifluorescence inverted microscope. A Leitz Planapo 63x, NA/1.4 oil
immersion lens and a modified Zeiss Axiomat inverted light microscope was used for
collecting three-dimensional light microscopy data sets (Agard, 1984; Agard and Sedat,
1983). The focal position, UV excitation shutter, and digital camera shutter of the
microscope were under computer control. The images generated were digitized directly
from the microscope image plane using a 14 bit, liquid nitrogen-cooled charge-coupled
device (CCD) digital camera (described in detail in Hiraoka et al., 1987; Agard et al.,
1989; Paddy et al., 1990). Three-dimensional data sets were collected as a series of
images separated by 0.5 mm along the horizontal optical sectioning axis (this value varies
with the depth of the cell). For double-labeling experiments consisting of different
fluorescence wavelengths, a complete 3-D data set at the first wavelength was collected,
the focus was returned to the starting focal position and the barrier filter was changed
using the computer control, then the second data set was collected (Hiraoka et al., 1991).
After data collection, each 3-D data set was corrected for stage and/or sample drift,
fluorescence photo-bleaching through the data set, and lamp intensity and/or shutter open

22
time variations. Following image correction, 3-D deconvolution corrected for the out-of-
focus contamination from each optical section. The images were displayed using an
integrated, multiple-windowed, mouse-driven display and Delta Vision software (Applied
Precision. Issaquah, WA).
Expression of exogenous transporter by transfection. The conditions for
transfection of the transporter cDNAs were optimized using human fibroblasts. The same
transfection protocol was used for Hela, PAEC, and HepG2 cells. Cells were plated onto
cover slips in 6-well trays 24 h before transfection. Optimal density for transfection was
within the range of 60-80% confluence for all cell lines used. Cells were transfected
using lipofectamine in Opti-MEM I reduced-serum medium according to the standard
protocol from Gibco Laboratories. For each transfection, 1 pg of a cDNA was added to
100 pi of Opti-MEM in a 17 x 100 mm polystyrene tube, and 4 pi of lipofectamine
reagent was mixed with 100 pi of Opti-MEM medium in a second polystyrene tube. The
solutions in each tube were combined and allowed to incubate for 30 min at room
temperature in order for cDNA-liposome complexes to form. During the incubation, cells
were washed two times quickly with PBS. For each transfection, 0.9 ml of serum- and
antibiotic-free MEM medium was mixed with the cDNA-liposome complexes and 1 ml
of the final transfection solution was applied to the cells. The cells were incubated with
the mixture for 3 h at 37C before washing two times with PBS and adding MEM
supplemented with antibiotics and 10% FBS. After 24 h, cells were fixed and labeled
according to the immunofluorescence protocol as described above. This lipofectamine
protocol above was compared to the liposome-mediated transfection protocols from either

23
Quiagen or Boehringer Mannheim and proved to be the least cytotoxic and provide the
greatest transfection efficiency (about 15-20%). To remove endotoxins from the cDNA
preps. cDNA in solution was mixed with 1% Triton X-l 14. vortexed, and chilled on ice
for 5 min. The sample was heated for 5 min at 37C, and centrifuged at 14,000 x g for 5
min before recovering the aqueous solution to be used in the transfection protocols.
Preparation of transporter cDNA-Green Fluorescent Protein constructs. The
pEGFP vectors from Clontech (Palo Alto, CA.) encode the Green Fluorescent Protein
(GFP) variants that fluoresce 35 times more intensely than wild-type and have been
codon-optimized for maximal translation efficiency in mammalian cells (Cormack et al.,
1996). These vectors are available in all three reading frames and contain 20 unique
restriction sites in the multiple cloning region to facilitate subcloning. The pEGFP(N3)
Protein Fusion Vector from Clontech was used to fuse the EAAT1 cDNA to the N-
terminus of EGFP(N3). The EAAT1 cDNA in pCDNA3 was obtained from Dr. Jeffrey
Rothstein's laboratory. PCR primers were designed to the 5 terminus beginning at the
ATG translation start site (ATGACTAAAAGCAATGGAGAAGAGC) and the 3
terminus ending at nucleotide 1680 (CATCTTGGTTTCACTGTCGATGG). Using 5 ng
of EAAT1 cDNA and 100 pmol of each primer, the entire coding region minus the stop
codon was amplified with Taq polymerase according to the manufacturers protocol
(InVitrogen, Carlsbad, CA). Amplification proceeded for 30 cycles of the following
conditions: denaturation at 94C for 1 min. annealing at 45C for 1 min, and extension at
72C for 1 min, with a final extension of 10 min. After the PCR product was obtained,
the 1680 base pair band was gel purified according to the protocol of Quiagen (Valencia,
CA) and cloned into the TA cloning vector, pCR 2.1 according to the manufacturer's

24
protocol (InVitrogen). A partial digest was performed using EcoRI restriction enzyme to
isolate the EAAT1 1680 base pair fragment. This fragment was then subcloned into the
pEGFP(N3) vector, at the EcoRI site, in order to place the EAAT1 cDNA before the
GFP. The 1680 base pair fragment was also cloned into the EcoRI site of the pEGFP(C3)
vector, which placed EAAT1 just before a stop codon at the C-terminus of the GFP.
A 2280 base pair fragment, including the entire coding sequence and stop codon,
of the CAT1 cDNA was cut out of pCDNA3 with Hindlll and EcoRI and subcloned
directly into the multiple cloning site of pEGFP(C3) at the Hindlll and EcoRI sites, thus
placing CAT1 at the C-terminal end of GFP.
Mutagenesis. Oligonucleotide-directed site-specific mutagenesis of the CAT1
cDNA was performed using the MORPH Mutagenesis kit from 5 Prime 3 Prime, Inc.
(Boulder, CO). CATMUT1 was constructed using the wild type GFP(C3)-CAT1 cDNA
as the template and the CAT1.1 mutant oligonucleotide, GGT GTT GAG GGA GCG
GGA CAG GCG GCT CTC CTC CCG GCT GGA GTC GAC CAC CTT CCG GCG.
This mutagenesis reaction resulted in the substitution of serine for cysteine at residues 20
and 30. CATMUT2 was constructed using the GFP(C3)-CAT1 cDNA as the template
and the CAT1.2 mutant oligonucleotide, (GCA TCT GCT GGC CCA GCC CGA GCA
GGT TTT TGG AGGCCA TTG TGC TGA GCG AAT CTG C). This reaction resulted
in the substitution of alanine for glycine at position 2. CATMUT3 was generated using
CATMUT1 as the template and the mutant oligonucleotide, CAT1.3 (GCA TCT GCT
GGC CCA GCC CGA GCA GGT TTT TGC AGG CCA TTG TGC TGA GCG ATT
CTG C). This reaction resulted in the substitution of alanine for glycine at position 2,

25
and serine for cysteine at positions 3, 20, and 30. Each mutagenic oligonucleotide above
was designed with an internal restriction site for analyzing the success of the mutagenesis
reaction. A Sal I restriction site was engineered into the CAT 1.1 mutagenic
oligonucleotide and the CAT1.2 and CAT1.3 mutagenic oligonucleotides were
constructed with an internal Ava I site. Prior to beginning the mutagenesis, 2.5 pg of the
oligonucleotide was 5' phosphorylated in a reaction using 5 pi of 10X T4 polynucleotide
kinase buffer, 25 U of T4 polynucleotide kinase, and 10 mM ATP. This reaction was
allowed to proceed at 37C for 1 h before being terminated by heating to 65C for 10 min.
The annealing procedure involved mixing 0.03 pmol of the target cDNA (GFP-CAT1),
2.0 pi of 10X MORPH annealing buffer, and 100 ng of phosphorylated mutagenic
oligonucleotide and heating the solution to 100C for 5 min to denature the double-
stranded DNA. At this point, solutions were either placed at room temperature for 30
minutes, or placed in a beaker of 70C water that was allowed to cool slowly to room
temperature. One procedure worked better than the other for specific oligonucleotides
and the choice was determined experimentally. For the replacement strand reaction, 8 pi
of 3.75X MORPH synthesis buffer, 3 U T4 DNA polymerase, and 4 U T4 DNA ligase
were added directly to the annealing reaction and incubated for 2 hr at 37C, then for 15
min at 85C to terminate the reaction. A 1:10 dilution of Dpn I restriction enzyme was
prepared and 1 pi was added to each mutagenesis reaction. The solution was incubated
for 30 min at 37C and then placed on ice for 5 min. The premise behind this digestion is
that Dpn I will specifically digest only double-stranded DNA in which both strands are
methylated, therefore, any double-stranded non-mutagenized target plasmid DNA will be

26
cleaved into small linear strands and will not be efficiently introduced and propagated
during the bacterial transformation. Following the digestion, the entire mutagenesis
reaction was added to 200 pi of E. coli MORPH mutS cells and incubated on ice for 20
min. The cells were heat-shocked at 42C for 2 min and spread on LB plates containing
30 pg/ml kanamycin for selection of the appropriate target plasmid. The mutS strain of
E. coli is used because it is deficient in DNA repair strand selection, therefore, it
randomly repairs either the mutant strand or the original template pair. This way, there is
a 50% chance that the mutant strand will be selected as correct and that sequence will be
propagated further. After selecting colonies and growing the bacteria in LB plus 30
pg/ml kanamycin, the plasmid DNA was isolated from several colonies with Quiagen
minipreps and digested using the enzyme specific to the engineered sequences to confirm
they contained the mutant strand (Sal I for CATMUT1 and Ava I for CATMUT2 and
CATMUT3). Large scale plasmid prep kits (manufactured by Quiagen) were used to
prepare the final mutant cDNAs. Each of the mutants was subjected to a series of
restriction digests to confirm the amino acid substitution(s) and to check for appropriate
sizes of the vectors and inserts.
Data analysis. Much of the data generated in the proposed experiments were
qualitative and required visual analysis. Each immunofluorescence experiment was
performed a minimum of three times to check for consistency among cell populations
plated on different days. In addition, normal and LPI cells of the same passage number
were assayed in parallel using the same reagents and antibody solutions. LPI cells from
several different patients were used for this study. Cells from the same patient were used

27
to check for reproducibility, then cells from a different patient were used to confirm that
the results are not unique to a single individual with the disease.

CHAPTER 3
DISTRIBUTION OF THE GLUTAMATE TRANSPORTERS
Introduction
Five plasma membrane proteins have been cloned that belong to the family of
transporters responsible for the Na'-dependent uptake of glutamate and aspartate in a
variety of tissues (see Chapter 1 for overview). These transporters correspond to the
high-affinity. Na~-dependent System XAG activity previously described in human
fibroblasts (Dall'Asta et al., 1983) and cultured liver cells (Makowske and Christensen.
1982). The glutamate transporters share approximately 50% amino acid identity, consist
of 6 or more transmembrane spanning domains, and contain two experimentally
identified N-linked glycosylation sites on the second extracellular loop (Conradt et al.,
1995). Glutamate uptake by these transport proteins is electrogenic. and coupled to the
co-transport of three Na~ ions and one H\ as well as the counter-transport of one Kr ion
(Zerangue and Kavanaugh. 1996).
Although similar in structure and substrate specificity, the glutamate/aspartate
transporters are differentially expressed and regulated. In the brain. EAAT1 (GLAST1)
and EAAT2 (GLT1) are localized to astroglia, whereas EAAT3 (EAAC1) is specific to
neurons. EAAT4 also has been localized to neurons and EAAT5 is specific to retinal
tissue. The family members also differ with respect to their transport properties. When
ooyctes were injected with EAAT5 cRNA, radiolabeled glutamate uptake was increased
28

29
by 2- to 10-fold over that of the control uninjected oocytes (Arriza et al., 1997).
However, this was significantly less than the transport observed with EAAT1, EAAT2.
and EAAT3-expressing oocytes, which were reported to transport 50-fold over the level
of uninjected oocytes (Klockner et al.. 1993; Kanai et al.. 1995). Whereas
electrophysiological studies have shown that both EAAT4 and EAAT5 transport
glutamate poorly, they possess much stronger chloride channel properties only weakly
present in the other glutamate transporters (Fairman et al., 1995; Arriza et al., 1997).
Several laboratories have documented the involvement of glutamate in cell migration and
differentiation (Pearce et al., 1987; Mattson et al., 1988), as well as neuronal and
astroglial proliferation. Osnat Bar-Peled et al. have shown that each of the glutamate
transporters has a specific and unique distribution during brain development (Osnat Bar-
Peled et al., 1997), suggesting that the transporters play multiple functional roles during
brain maturation.
Three amino acid residues in the C-terminal sequence, the region of greatest
homology, have been identified as essential for glutamate transport activity (Conradt and
Stoffel, 1995; Pines et al., 1995). Using site-directed mutagenesis. Pines and coworkers
determined that aspartate 398, glutamate 404, and aspartate 470 are critical for EAAT2
activity, and glutamate 404 may contribute to substrate specificity (Pines et al., 1995).
Both aspartate 398 and 470 appear to be involved in transporter activity, rather than
stability or trafficking, and even the conservative replacement of glutamate abolishes
transport activity. Zhang et al. showed, using site-directed mutagenesis, that histidine
326 is required for glutamate transport by EAAT2, and probably contributes to the proton
translocation mechanism that accompanies the Na and K7-coupled transport activity

30
(Zhang et al., 1994). Conradt and Stoffel performed similar mutagenesis studies using
the EAAT1 cDNA (Conradt and Stoffel. 1995). When they mutated the conserved
arginine 122. arginine 280. arginine 479. and tyrosine 405, they lost glutamate transport
activity with the tyrosine 405 and arginine 479 mutants. The arginine 122 and arginine
280 mutants appeared to increase the K, of EAAT1 for aspartate, but had no effect on the
intrinsic properties or kinetics of glutamate transport. They proposed from their studies
that the hydroxyl group on tyrosine 405 and the positive charge on arginine 479 may
contribute to the binding of the acidic glutamate substrate.
Although most of the research involving the glutamate transporters has been
confined to the brain. L-glutamate is crucial to several biochemical pathways of
peripheral tissues as well (i.e., ammonia detoxification and gluconeogenesis). Several
laboratories have independently shown, by mRNA and protein analysis, that tissues other
than the brain express one or more of the transporter isoforms. EAAT3 is the most
ubiquitous of the glutamate transporters and is detected in kidney, small intestine, liver,
heart, lung, skeletal muscle, and placenta (Kanai and Hediger, 1992; Matthews et al.,
1998). Glutamate and asparate are almost completely reabsorbed from the glomerular
filtrate by the abundantly expressed EAAT3 transporter in the renal tubules (Silbemagl,
1983; Shayakul et al., 1998). EAAT1 is expressed in heart, lung, skeletal muscle, retinal
glia, and placenta (Kawakami et al., 1994; Arriza et al., 1994). and both EAAT2 and
EAAT4 have also been detected in placental tissue (Matthews et al., 1998). Preliminary
data from our laboratory (Tessmann, unpublished results) also suggest that EAAT1,
EAAT2, and EAAT3 are expressed in human fibroblasts.

31
L-Glutamate and L-aspartate are important nutritional substances that contribute
to a variety of biochemical pathways in the brain and peripheral tissues. Specialized
glutamatergic neurons in the brain produce and store glutamate, the major excitatory
neurotransmitter in the mammalian central nervous system, until it is released into the
synaptic cleft in response to different stimuli. Intracellular concentrations of glutamate in
the brain reach approximately 10 mM, with the highest concentrations at nerve terminals
(Shupliakov et al., 1992; Storm-Mathisen et ah. 1992). Extracellular levels of glutamate
are carefully maintained below 3 pM, except during neurotransmission of a signal when
concentrations may reach between 1-2 mM (Nicholls. 1993; Clements et ah. 1992).
Members of the EAAT family of transporters, primarily the glial-specific EAAT1 and
EAAT2 (Rothstein et ah, 1996), are responsible for the high-affinity Na-dependent
transport of glutamate out of the synaptic cleft against a thousand-fold concentration
gradient. This carefully regulated transport activity is crucial for maintaining glutamate
concentrations below the level that is toxic to neurons.
Glutamatergic transmission is believed to contribute to normal brain activities
such as learning and memory (Bliss et ah, 1993), however, elevated levels of extracellular
glutamate are neurotoxic and can lead to several neuro-degenerative diseases such as
Amyotrophic Lateral Sclerosis (ALS), Huntingtons disease, and probably Alzheimers.
In 1992. Rothstein et ah showed that brain and spinal cord samples taken from the
autopsies of ALS patients revealed a reduced level of glutamate uptake (Rothstein et al.,
1992). More recently it was determined that the decrease in glutamate uptake in some
cases of sporadic ALS is due to the selective loss of the astroglial EAAT2 glutamate

32
transporter (Rothstein et al.. 1995). It has recently been proposed that the down-
regulation of EAAT2 results from a defect in mRNA processing. Lin et al. has shown
that due to defective mRNA splicing events such as intron-retention or exon-skipping,
multiple abnormal EAAT2 mRNA species are produced in the affected areas of the brain
in ALS patients (Lin et al.. 1998). In vitro expression studies suggested that the protein
products of the aberrant mRNAs had decreased transport activity because they were
degraded rapidly, or perhaps, had a dominant-negative effect on the normal EAAT2
protein. The abnormal mRNA species were not present in regions of the brain that were
unaffected in ALS patients, or in the non-neurologic disease controls.
Decreased glutamate transporter activity in the frontal, parietal, and temporal
cortex has been implicated in the neurodegeneration that occurs in Alzheimer disease
(Scott et al.. 1995; Cowbum et al.. 1988). Like ALS. mRNA levels of EAAT1. EAAT2.
and EAAT3 were normal in the frontal cortex, however, immunoblot analysis detected
about 30% less EAAT2 protein (Li et al.. 1997). On the other hand, schizophrenia and
other psychoses are thought to result partially from glutamatergic hypofunction. a
condition that occurs following excessive glutamate uptake (Carlsson and Carlsson.
1990). Therefore, the mechanism by which glutamate is cleared from the synaptic cleft
must be tightly regulated in order to prevent neuronal damage or malfunction.
The roles that the individual transporters play in normal synaptic clearance and
neurotoxicity are unclear because subtype-specific inhibitors are not available. This has
led a number of laboratories to study the functions of these specific transporters using
antisense oligonucleotides or knockout mice. Results from antisense studies indicated
that a loss of EAAT1 and EAAT2 resulted in elevated extracellular glutamate

33
concentrations as well as neurodegeneration and progressive paralysis (Rothstein et al..
1996). Antisense oligonucleotides to EAAT1. EAAT2. and EAAT3 were administered
intraventricularly to male Sprague-Dawley rats for 7-10 days. Within 3 days, the animals
that were treated with the EAAT1 and EAAT2 antisense oligonucleotides began to show
evidence of progressive motor degeneration, and by day 8. their hindlegs were paralyzed.
When extracellular levels of glutamate were measured with microdialvsis probes in the
ipsilateral striatum of treated rats, extracellular glutamate concentrations were elevated by
32-fold in EAAT2 antisense rats, and by 13-fold in EAAT1 antisense rats. Loss of
EAAT3 did not elevate extracellular glutamate levels in rats treated with EAAT3
antisense oligonucleotides. These animals did. however, experience epileptic seizures and
slightly impaired motor skills in some cases. These data support the findings that
EAAT1 and EAAT2 play a crucial role in synaptic clearance of glutamate, however, the
role of EAAT3 in preventing neurological damage seems to be unclear due to conflicting
data.
Evidence suggests that EAAT2 is responsible for the greatest amount of cerebral
glutamate transport (Robinson, 1991). When EAAT2 was knocked out by homologous
recombination, levels of residual glutamate increased in the brain and the mice suffered
lethal spontaneous seizures (Tanaka, 1997). Both the antisense and knockout studies are
consistent with the data documenting the loss of EAAT2 as being the major cause of the
motor neuron degeneration that plagues patients with ALS (Rothstein. 1995).
Conversely, when EAAT3 (EAAC1) knockout mice were produced, no
neurodegeneration was observed (Peghini and Stoffel, 1997). Instead, the mice
developed dicarboxylic aminoacidurea, analogous to an inborn error of glutamate and

34
aspartate transport across epithelial cells of the kidney and intestine. In some cases,
mental retardation or neurological abnormalities are also symptoms of the disease,
however, there was no evidence of neurological damage in the EAAT3 knockout mice.
The dicarboxylic aminoacidurea acquired by the knockout mice is explained by the fact
that EAAT3 is strongly expressed in the kidney and is responsible for tubular
reabsorption of glutamate and aspartate from the glomerular filtrate.
Based on their accepted function of Na'-dependent glutamate/aspartate transport
and the membrane spanning structure of these amino acid transporters, we predicted that
the EAAT transporters would be primarily localized to the plasma membrane, with some
intracellular pools involved in biosynthesis or recycling. However, this could not be
assumed based on the recent detection of several plasma membrane proteins in the
nuclear membrane and matrix. P-glycoprotein. a 170 kDa protein with 12 membrane
spanning domains, has been implicated in conferring multi-drug resistance in cancer cells
(Juliano and Ling. 1976). It accomplishes this by actively exporting a wide-range of
chemotherapeutic agents out of the cell in an ATP-dependent mechanism. Recently. P-
glycoprotein was reported to reside in the nuclear membrane and matrix as well as on the
plasma membrane (Baldini. 1995). Also, a number of laboratories have reported the
presence of different growth factor receptors associated with the nucleus (Stachowiak et
al., 1996; Podlecki et ai 1987; Rakowicz-Szulczvnska et al., 1989). Traditionally, these
receptors were believed to reside on the plasma membrane and transmit a signal from the
extracellular environment to the cytosol upon ligand binding. However, the fibroblast,
insulin, and epidermal growth factor receptors are all integral membrane proteins that
have been localized to the nucleus.

35
In a study of the activity and localization of the glutamate transporters in day 14
and day 20 rat placenta. Matthews et al. detected EAAT1 in the nuclei of the maternal
decidua and placental trophoblast (Matthews et al.. 1998). In other cell types of the
placenta, however. EAAT1 was detected on the plasma membrane and in one or more
intracellular vesicle populations. EAAT1. EAAT2, and EAAT3 mRNA and protein
expression were increased in the day 20 placenta, and the expression patterns were cell-
type specific for each isoform. Neither EAAT2 nor EAAT3 was observed in the nuclei of
any of the cell types examined by immunohistochemistry. Although EAAT4 mRNA was
identified in both day 14 and day 20 placenta. EAAT4 protein was not detectable in either
sample by immunoblot or immunohistochemistry.
Over the past two decades, extensive research has been performed to determine
the ionic requirements, substrate specificity and kinetic parameters of the glutamate
transporters, however, there is almost no information regarding intracellular pools,
trafficking through compartments, or protein arrangement on the plasma membrane.
One of my intentions for this project was to answer some basic questions regarding the
cellular and subcellular distribution of the different isoforms. Using sequence-specific
antibodies generated against the individual EAAT transporters. I could explore the
expression of the isoforms in cell lines other than the brain. In addition, I could
determine if the different EAAT family members display a similar or distinct staining
pattern in regard to abundance and distribution of protein. This chapter describes the
results obtained from staining human fibroblasts with antibodies against the glutamate
transporters EAAT1 and EAAT3. More extensive work involving co-localization
experiments in several cell lines was performed using EAAT1 antibodies from several

36
independent sources. The EAAT1 cDNA was used to create a fusion protein with the
green fluorescent protein (GFP) and distribution of the transporter was examined
following transient expression of this EAAT1-GFP fusion protein. These experiments
were performed in combination with immunofluorescence and immunoblotting using four
different antibodies specific to EAAT1.
Methods
Glutamate transporter antibody production. The polyclonal EAAT3 glutamate
transporter antibody was raised in rabbit against an EAAT3-maltose binding protein that
was constructed by Dr. Marc Malandro using the C-terminal 120 amino acids of the rat
EAAT3 (EAAC1) sequence. The glutamate transporter antibody designated EAAT1-R
was obtained from Dr. Jeffrey Rothstein at Johns Hopkins University. EAAT1-R is a
polyclonal antibody that was raised in rabbit against the amino acid residues 3-17
(KSNGEEPRMGSRMGR) at the N-terminus of the human EAAT1 protein (Ginsberg et
al., 1995). The EAAT1 glutamate transporter antibody designated EAAT1-S was
obtained from Dr. Wilhelm Stoffel at the University of Cologne (Cologne, Germany).
This polyclonal antibody was generated in rabbit against a peptide consisting of amino
acid residues 24-40 (KRTLLAKKKVQNITKED) at the N-terminus of the rat EAAT1
sequence (Wahle and Stoffel, 1996). The EAAT1-C polyclonal antibody was generated
in guinea pig, by Chemicon International, Inc. (Temecula, CA), against the rat C-terminal
peptide, QLIAQDNEPEKPVADSETKM (Storck et al., 1992). The polyclonal EAAT1-
D antibody was purchased from a-Diagnostics International (San Antonio, TX). This
antibody was raised in rabbit against a peptide (amino acids 504-518,

37
NRDVEMGNSVIEENE) from the C-terminus of the rat EAAT1 sequence (Rothstein et
al., 1995).
Cell fractionation. Media was aspirated from human fibroblasts grown to 80-90%
confluence in 150-mm dishes. Cells were washed three times with 10 ml ice cold PBS or
SEB (85.6 g/L sucrose. 0.76 g/L EGTA. 2.38 g/L Hepes, pH 7.5). then scraped and
collected in a plastic 50 ml centrifuge tube. A total of 15 ml of SEB with 0.5 mM
phenylmethyl sulfonyl fluoride (PMSF) and 1 pl/ml protease inhibitors (1 pg/ml antipain
and leupeptin. and 100 KIU/ml aprotinin) per 150-mm dish was used to scrape the cells.
Using a refrigerated table-top centrifuge, cells were spun at 300 x g for 5-10 minutes,
supernatant discarded, and cell pellets resuspended in 15 ml of SEB + PMSF and protease
inhibitors. Cell suspension was poured into an ice-cold nitrogen bomb (Parr Instrument
Company, Moline. IL) and allowed to equilibrate at 200 psi for 10 min before lysis by
rapid release. Resultant homogenate suspension was collected in a centrifuge tube and
checked for unbroken cells under a light microscope. Homogenate was centrifuged at
300 x g for 10 min as described above, and the pellet was saved as the nuclear fraction.
The 300 x g supernatant was spun in an ultracentrifuge at 15.000 x g (14.500 rpm in a
Beckman 60TI rotor) for 30 min and the pellet saved as a crude plasma membrane-
enriched fraction. The supernatant from the 15,000 x g spin was centrifuged in the
ultracentrifuge at 100.000 x g (37.500 rpm in a Beckman 60TI rotor) for 60 min and the
pellet was saved as the total intracellular membrane fraction.
SDS-PAGE and immunoblot analysis. Samples from the cell fractionation
procedure were initially dissolved in 0.2 N NaOH/O.2% SDS, and then further diluted to

38
0.5-1.0 pg/pl in sample dilution buffer and 10-20 pg protein per lane were subjected to
one-dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE)
(Laemmli. 1970; Chiles et al., 1987). The proteins were transferred at 299 mAmps for 18
h to a nitrocellulose membrane (Chiles et al., 1987). The nitrocellulose membrane was
blocked with 5% non-fat dry milk (NFDM) at room temperature for 1.5 h. rinsed with
TBS/TWEEN (30 mM Tris base, 150 mM NaCl, 0.1% Tween 20, pH 7.6), then incubated
with primary antibody (summarized below) prepared in TBS/TWEEN for 1-2 h at room
temperature. Two quick rinses, one 15 min rinse, and two 5 min rinses with
TBS/TWEEN were followed by incubation of the nitrocellulose in the secondary
antibody, prepared in TBS/TWEEN, for 1-2 h at room temperature. The EAAT1-R
antibody was used at 1:200-1:1000 dilutions, the EAAT1-S antibody was used at a 1:50
dilution, and both were visualized using a 1:5000 dilution of either a donkey or goat anti
rabbit IgG conjugated to horseradish peroxidase. The EAAT1-C antibody was used at a
dilution of 1:100 and detected with a 1:10,000 dilution of goat anti-guinea pig IgG
conjugated to horseradish peroxidase. After washing the nitrocellulose blot six times for
5 min each with TBS/TWEEN, the blot was incubated in 6 ml of a 1:1 mixture of the
Enhanced Chemiluminescence (ECL) reagents (Pierce, Rockford. IL) for 1 min, drained,
wrapped in plastic, and immediately exposed to film.
Antibodies for immunoblotting and immunofluorescence. The antibodies used for
immunoblotting (described above) and immunofluorescence (described in Chapter 2) are
summarized in Table 3-1. Unless otherwise stated, EAAT1 and EAAT3 antibodies were
detected during immunofluorescence assays using a 1:200 dilution of goat anti-rabbit IgG

39
conjugated to fluorescein isothiocyanate (FITC). For double-labeling experiments,
monoclonal antibodies against the organelle-specific proteins were detected using a 1:200
dilution of goat anti-mouse IgG linked to Texas Red. The results of all
immunofluorescence experiments in this chapter were analyzed using deconvolution
microscopy (described
in the Methods Chapter).
Table 3-1
Antibodies against Glutamate Transporters
Name
Host
Source
Dilutions
jp* IB**
EAAT3
rabbit
Dr. Michael Kilberg.
Univ. of Florida
1:200
--
EAAT1-R
rabbit
Dr. Jeffrey Rothstein.
Johns Hopkins
1:50
1:200-
1:1000
EAAT1-S
rabbit
Dr. Wilhelm Stoffel,
Univ. of Cologne, Germany
1:50
1:50
EAAT1-C
guinea pig
Chemicon International.
Temecula, CA
1:1000
1:100
EAAT1-D
rabbit
a-Diagnostics,
San Antonio, TX
1:50

KDEL
mouse
Dr. David Vaux,
EMBL
1:100

Transferrin receptor
mouse
Zymed,
San Francisco, CA
1:5

414
mouse
Dr. John Aris,
Univ. of Florida
1:10

D77
mouse
Dr. John Aris,
Univ. of Florida
1:50

*IF = immunofluorescence assay
**IB = immunoblot
Results
Localization of endogenous EAAT3 glutamate transporter in human fibroblasts by
immunofluorescence. Immunofluorescence assays were performed on human fibroblasts
using antibodies generated against the glutamate transporters, EAAT1 and EAAT3.

40
Preliminary experiments with antibodies specific for EAAT2 and EAAT4 revealed no
detectable staining (data not shown). These experiments were intended to provide some
information concerning the similarities and differences in cellular distribution of amino
acid transporters belonging to the same gene family, and perhaps gain insight into why
multiple transporters with nearly identical kinetics are expressed in the same cell. For all
extracellular labeling experiments, fibroblasts were fixed with a 4% paraformaldehyde
solution and stained according to the immunofluorescence protocol described in the
Methods Chapter.
A 1:200 dilution of the anti-EAAT3 antibody stained the cells in clusters rather
than diffusely labeling the entire cell surface (Figure 3-1 A). This pattern of transporter
"patching" was similar to the pattern observed using the CAT1 arginine transporter
antibody (described in Chapter 5. Figure 5-1 A and B). Labeling by the anti-EAAT3
transporter antibody was completely inhibited by preadsorption of the antibody with 50
pg/ml of the corresponding peptide antigen for 12 h at 4C (Figure 3-1 B). The
intracellular distribution of the EAAT3 glutamate transporter was examined in human
fibroblasts following either -20C MeOH fixation, or 4% paraformaldehyde fixation in
combination with 0.1% Triton X-100 membrane permeabilization. Under both
conditions, the EAAT3 antibody labeled small vesicles throughout the cytoplasm (Figure
3-2 A). Double-labeling MeOH-fixed fibroblasts with antibodies against EAAT3 (1:200
dilution) and KDEL (1:100 dilution), a common epitope of resident proteins of the
endoplasmic reticulum (ER), showed very little co-localization (Figure 3-2 B). Also,
only a small amount of co-localization was detected when MeOFi-fixed fibroblasts were

41
double-labeled with a 1:200 dilution of the EAAT3 antibody and a 1:5 dilution of the
transferrin receptor (Figure 3-2 C). indicating that ven little of the EAAT3 glutamate
transporter is involved in recycling under standard culture conditions.
Localization of endogenous EAAT1 glutamate transporter in human fibroblasts bv
immunofluorescence. Initial experiments were performed using the EAAT1-R antibody
generated in Dr. Jeffrey Rothstein's laboratory against an N-terminal peptide sequence.
This antibody will be referred to as EAAT1-R (see Table 3-1 for details on EAAT1
antibodies). Unlike the "patches" observed on human fibroblasts with the anti-EAAT3
antibody, the anti-EAATl-R antibody did not detect any protein on the surface of
paraformaldehyde-fixed fibroblasts even at a 1:25 antibody dilution. When fibroblasts
were fixed with -20C MeOH and stained for intracellular EAAT1, using a 1:50 dilution
of the anti-EAATl-R antibody, the nuclear membrane and nuclear matrix were the
primary structures detected (Figure 3-3 A). Small vesicles throughout the cytoplasm,
resembling the vesicles labeled with the EAAT3 antibody, were also apparent. Staining
of the anti-EAATl-R transporter antibody was completely inhibited by preadsorption of
the antibody with 50 pg/ml of the corresponding peptide antigen for 12 h at 4C (Figure
3-3 B). To confirm that the EAAT1-R antibody was staining the nuclear membrane,
MeOH-fixed fibroblasts were double-labeled with antibodies against EAAT1-R (1:50
dilution) and 414(1:10 dilution) (Figure 3-3 C). The latter is an antibody generated
against an epitope shared by several proteins of the nuclear pore complex (Davis and
Blobel, 1986). There was significant co-localization of the EAAT1-R and 414
antibodies, strongly suggesting that EAAT1-R was labeling the nuclear membrane. On

42
the other hand, very little co-localization was detected when EAAT1-R and KDEL
antibodies were used (data not shown), indicating that the fluorescence was not likely due
to staining of ER components surrounding the nucleus. A D77 antibody, generated
against the yeast nucleolar protein, fibrillarin (Noplp), was used at a dilution of 1:50 in
double-labeling experiments with EAAT1-R antibody (1:50 dilution) to determine if the
nuclear staining was nucleolar (Aris and Blobel. 1988). The D77 antibody detected three
or four nucleoli in each cell, however, no co-localization with EAAT1-R was observed
(Figure 3-3 D).
The nuclear membrane staining of human fibroblasts with an antibody against a
known plasma membrane amino acid transporter (EAAT1) was unexpected. In an effort
to confirm the nuclear localization of the EAAT1-R antibody, a second EAAT1 antibody
was obtained from Dr. Wilhelm Stoffel (University of Cologne. Germany) and will be
referred to as EAAT1-S (Table 3-1). Although EAAT1-S was also generated against an
N-terminal peptide, there was no overlap with the amino acid sequence that EAAT1-R
was raised against. Therefore, EAAT1-S provided an additional antibody in case the
EAAT1-R antibody was cross-reacting to an unknown protein with homology to the
peptide used to generate EAAT1-R. When MeOH-fixed human fibroblasts were
incubated with a 1:50 dilution of the anti-EAATl-S antibody, no nuclear membrane or
matrix staining was observed (Figure 3-4 A). Only small vesicles throughout the
cytoplasm were detected.
With the apparent contradicting results obtained from immunofluorescence
experiments using the EAAT1-R and EAAT1-S antibodies, two additional EAAT1

43
antibodies were purchased from Chemicon (EAAT1-C) and a-Diagnostics (EAAT1-D).
Both antibodies were incubated with MeOH-fixed human fibroblasts according to the
same procedures used for the EAAT1-R and EAAT1-S antibodies. Nuclear staining was
observed using a 1:1000 dilution of the EAAT1-C antibody (Figure 3-4 B), detected with
a 1:400 dilution of goat anti-guinea pig IgG linked to Cy3. but that labeling associated
with the nuclear membrane was fainter than with the EAAT1-R. and there also was
significant staining of a cellular vesicle population. Thus, the pattern of fluorescence
generated by the EAAT1-C antibody appeared to be a mixture of the results obtained
from the EAAT1-R and EAAT1-S antibodies. No specific or background staining was
detected when human fibroblasts were incubated with EAAT1-C priman' or anti-guinea
pig secondary antibodies alone. However, attempts to enhance the EAAT1-C nuclear-
specific staining by centrifuging (5 min at 13.000 x g) or filtering the EAAT1-C antibody
through a 2 pm pore syringe prior to use failed. A 1:50 dilution of the EAAT1 -D
antibody resulted in a uniform, diffuse cytosolic labeling with no specific staining of the
nuclear membrane, cytoplasmic vesicles, or other cellular structures (data not shown).
Localization of the EAAT1 glutamate transporter in other cell types. To
determine whether or not the nuclear localization of EAAT1 was unique to fibroblasts,
several other cell types were labeled with both the EAAT1-R and EAAT1-S antibodies.
Hela. HepG2 (human hepatoma), and pulmonary artery endothelial cells (PAEC) were
each fixed with -20C MeOH and stained with 1:50 dilutions of both EAAT1-R and
EAAT1-S according to the immunofluorescence assay described in the Methods Chapter.
The EAAT1-R antibody labeled the nuclei of both Hela (Figure 3-5 A) and HepG2 (data

44
not shown), however, no nuclear staining was observed in the PAEC (Figure 3-9 A).
Instead, the EAAT1-R antibody stained small vesicles throughout the cytosol of the
PAEC. with slightly more fluorescence concentrated in the perinuclear region. As was
the case for the human fibroblasts, no nuclear staining was detected with the EAAT1-S
antibody in any of the cell lines. In Hela. HepG2, and PAEC, the EAAT1-S antibody
detected only small vesicles of unknown origin in the cytosol. Figure 3-5 (A and B)
shows representative Hela cells stained with the EAAT1-R and EAAT1-S antibodies. _
Identification of EAAT1 glutamate transporter by immunoblot analysis. Total
intracellular membrane (100,000 x g), crude plasma membrane-enriched (15,000 x g),
and nuclear fractions (300 x g) from human fibroblasts, HepG2. and PAEC were isolated,
subjected to SDS-PAGE, and transferred to nitrocellulose membranes for
immunoblotting. The primary antibodies were detected using 1:2.500 to 1:10.000
dilutions of goat anti-rabbit IgG (EAAT1-R and EAAT1-S) or goat anti-guinea pig IgG
(EAAT1-C) conjugated to horseradish peroxidase (described in the Methods Section of
this chapter). When the 300 x g nuclear fraction from human fibroblasts was incubated
with a 1:1000 dilution of EAAT1 -R (Figure 3-6), a strong band was detected at
approximately 70-75 kDa, corresponding to the molecular mass of the monomeric
EAAT1 protein (Tessmann and Kilberg, unpublished data). Less intense bands were
observed at higher molecular masses, including a light band at approximately 180 kDa,
which corresponds to the molecular mass of a putative EAAT1 trimer. EAAT1 was also
detected in the 100,000 x g membrane fraction, however, the protein was primarily
detected at approximately 180 kDa, which is probably the trimeric form. Our laboratory
observes that the more manipulation of the sample, such as the additional centrifugation

45
steps required to obtain the 100.000 x g pellet, the greater the shift from monomer to
dimer/trimer forms of the transporter. Also, detection of the higher molecular mass
species of EAAT1 is consistent with published data from other laboratories documenting
the formation of glutamate transporter homomultimers (Haugeto et al., 1996).
When HepG2 cells were subjected to cell fractionation, SDS-PAGE, and
immunoblotting (see Methods section above), a 1:200 dilution of EAAT1-R was detected
in the 300 x g fraction in both the monomeric (approximately 70 kDa) and higher
molecular mass form (approximately 180 kDa) (Pappas and Kilberg, unpublished data).
EAAT1-R immunoreactivity was detected in the 180 kDa form in the 15.000 x g and
100,000 x g fractions, as well as two smaller species (approximately 70 and 75 kDa) in
the 100.000 x g fractions (data not shown).
Cell fractionation. SDS-PAGE. and immunoblotting were also performed using
PAEC according to the protocols in the Methods Section of this chapter. To determine if
the EAAT1 immunoreactivity in the nuclear (300 x g) fraction was a result of
contamination by other membranes or unbroken cells, a 1:100 dilution of A-l A5 ((3-
integrin) antibody was incubated with each of the fractions (data not shown). The A-1A5
antibody detects one or more p-integrin species (depending on the cell type) that migrate
at 210, 165, and 130 kDa (Hemler et al., 1984), and are specific for the plasma
membrane. I have confirmed the plasma membrane specificity of the antibody by SDS-
PAGE and immunoblot analysis (data not shown). Bands of approximately 165 kDa
were detected by the P-integrin antibody in the 300 x g fraction as well as the 15,000 x g
fraction of PAEC. Very little protein was observed in the 100,000 x g fraction. Although

46
the majority of the P-integrin labeling appeared in the plasma membrane-enriched
fraction, the band in the nuclear fraction suggests that the 300 x g fraction is
contaminated with plasma membrane, or more likely unbroken cells. When a 1:100
dilution of EAAT1-C was incubated with PAEC fractions, a band at approximately 75
kDa was detected in all three fractions in almost equal amounts (data not shown).
Transfection of EAAT1-GFP constructs in human fibroblasts. Two fusion
constructs between EAAT1 and green fluorescent protein (GFP) were generated to
compare the localization of the expressed transporter with the endogenous EAAT1
immunofluorescence experiments. The EAAT1-GFP(N3) fusion protein was constructed
with the GFP tag at the C-terminal end of EAAT1. and the GFP(C3)-EAAT1 was
constructed with GFP at the N-terminal end. The GFP tag was attached to either the N-
or C-terminus of EAAT1 to ensure that the location of the GFP did not interfere with
normal transporter trafficking (see Methods Chapter for details). Both of these constructs
were expressed in human fibroblasts using the lipofectamine transfection procedure
described in the Methods Chapter. After 24 hours of expression, cells were fixed with
-20C MeOH. mounted on glass slides with Fluoromount-G, and visualized using an
FITC filter on the deconvolution microscope, fluman fibroblasts transfected with the
GFP vector only (Figure 3-7 A) showed diffuse fluorescence throughout the entire cell,
whereas cells transfected with either EAAT1-GFP(N3) (Figure 3-7 B) or GFP(C3)-
EAAT1 (Figure 3-7 C) showed distinct vesicles in the cytoplasm as well as staining of
the plasma membrane. After analysis of three separate experiments, it was concluded that
there was no difference in the fluorescent patterns generated by the two different GFP

47
fusion proteins. Although some perinuclear fluorescence was observed, probably Golgi
localization, the pattern was clearly distinct from the nuclear membrane labeling observed
with the EAAT1-R antibody.
Co-localization of EAAT1-GFPN3) and the endogenous EAAT1 in human
fibroblasts. The fluorescence pattern observed in the EAATl-GFP(N3)-transfected
fibroblasts was consistent with the pattern of staining by the EAAT1-S antibody, that is
fluorescence was associated with a population of small vesicles scattered throughout the
cytoplasm. However, to test the specificity of both the EAAT1-S and EAAT1-R
antibodies for the exogenous transporter, human fibroblasts were transfected with the
EAAT1-GFP(N3) fusion protein, fixed with -20C MeOH, and stained with either the
EAAT1-R (Figure 3-8 A) or EAAT1-S (Figure 3-8 B) antibody. The EAAT1-S and
EAAT1-R primary antibodies were detected using a 1:200 dilution of goat anti-rabbit IgG
conjugated to Texas Red. A significant amount of co-localization was observed with the
EAAT1-GFP(N3) fusion protein and EAAT1-S antibody, however, little overlap was
detected when the EAAT1-R antibody was used. In the fibroblasts stained with EAAT1-
S there were separate pools of Texas Red-labeled vesicles that did not overlap with
EAAT1-GFP(N3). This could be due to the EAAT1-S antibody recognizing endogenous
EAAT1 which, for unknown reasons, was localized in a compartment lacking the
expressed fusion protein. Conversely, there was also a population of vesicles containing
EAAT1-GFP(N3) that were not stained with EAAT1-R. This latter result can be
explained by two vesicle populations or by the fact that the antibody-antigen binding
efficiency is less than 100% for the immunofluorescence assay. These results strongly

48
suggest that the EAAT1-S, but not the EAAT1-R, antibody recognizes the expressed
EAAT1-GFP(N3) protein in the human fibroblasts.
Co-localization of EAAT1-GFP(N3) with the EAAT1 endogenous in PAEC. As
an extension of the experiments in the last section. PAEC were transfected with the
EAAT1-GFP(N3). fixed with -20C MeOH. and stained with either the EAAT1-R (Figure
3-9 A) or EAAT1-S (Figure 3-9 B) antibody. Primary antibodies were detected using a
1:200 dilution of goat anti-rabbit IgG linked to Texas Red. PAEC were selected because
they were the only cell type tested that did not show nuclear localization of EAAT1 in the
initial experiments with the EAAT1-R antibody. As observed with the human
fibroblasts, there was significant co-localization between the EAAT1-GFP(N3) fusion
protein and the EAAT1-S antibody in PAEC. Although the fluorescent patterns of the
EAAT1-R antibody and EAAT1-GFP(N3) were both punctate and cytoplasmic, there was
little overlap between the two stains.
Discussion
This project began with the observation that an antibody (EAAT1-R) against a
known plasma membrane transporter (EAAT1) labeled the nucleus of human fibroblasts.
The EAAT3 isoform, on the other hand, was detected on the plasma membrane and in
vesicles throughout the cytoplasm. The clustering observed using the anti-EAAT3
antibody is consistent with the report by Davis et al. showing punctate fluorescence
throughout the cytoplasm with clustering at the cell surface (Davis et al., 1998).
Although the EAAT1 nuclear localization was unexpected, it was not unprecedented. As
discussed in the Introduction to this chapter, Matthews et al. detected nuclear staining

49
with the EAAT1-R antibody in the maternal decidua and placental trophoblast cells of rat
placenta using immunocytochemistry (Matthews et al.. 1998). In other cells of the
placenta, no nuclear localization was detected, but rather, EAAT1 was distributed
throughout the cytoplasm and on the plasma membrane. Other plasma membrane
proteins, such as P-glycoprotein and various growth factor receptors, have also been
detected in the nucleus and nuclear membrane by immunohistochemistry (Stachowiak et
al., 1996; Baldini et al 1995).
There are various explanations for the data presented above. First, the EAAT1
glutamate transporter may reside in the nucleus and provide a function that is different
from that of the plasma membrane nutrient transporter. If EAAT1 is serving as an amino
acid transporter in the nucleus, it is unlikely to be mechanistically similar to that on the
cell membrane due to the lack of a Na* gradient across the nuclear membrane. However,
the glutamate transporters also carry other ions such as, K* and H\ and the isoforms.
EAAT4 and EAAT5, have intrinsic chloride channel properties. Perhaps, ion transfer is
the primary function within the nuclear membrane. On the other hand, the lack of nuclear
localization observed with the EAAT1-GFP(N3) and GFP(C3)-EAAT1 fusion proteins,
as well as with both the EAAT1-S and EAAT1-D antibodies, presents evidence against
nuclear localization of the native EAAT1 protein. Although no nuclear localization
signaling sequence has been identified in EAAT1, there are hundreds of nuclear
localization sequence variations which are not highly conserved (Boulikas, 1996).
Another alternative is that the protein recognized by the EAAT1-R antibody
might prove to be a nuclear isoform of EAAT1 with a conserved sequence at the N-
terminus. The fact that the EAAT1-S antibody, generated against a different peptide

50
sequence, did not stain the nucleus would be consistent with the hypothesis that the
EAAT1-R antibody recognized a different, yet related, protein to EAAT1. Also, if a
truncated form of EAAT1 resides in the nucleus, then it is possible that it would be
recognized by the EAAT1-R but not the EAAT1-S antibody. This later explanation,
however, is not supported by the data from the immunoblot analysis in which the
EAAT1-R antibody recognized a nuclear protein corresponding to the molecular weight
of the intact EAAT1 protein.
The third explanation is that the EAAT1-R antibody is detecting a protein that is
completely different from the EAAT1 glutamate transporter, but shares a few amino acids
in common. The peptide inhibition indicates that the EAAT1-R antibody interaction is
specific, therefore, the unrelated nuclear protein would most likely have at least one
homologous epitope to EAAT1. BLAST database searches did not identify any proteins
other than EAAT/ASCT family members, with significant homology to EAAT1, but such
searches do not always reveal small stretches of common sequence. Also, it would be
quite coincidental that the unknown protein has the same molecular mass as EAAT1 on
an SDS-PAGE gel. The fact that the EAAT1-R antibody showed very little co
localization with the EAAT1-GFP(N3) fusion protein is puzzling. It is surprising that the
EAAT1-R antibody would recognize an unrelated or homologous protein, and not detect
the expressed EAAT1-GFP(N3). However, the fact that EAAT1-S recognized the
EAAT1-GFP(N3), suggests that the EAAT1-S antibody is specific for EAAT1, and does
not recognize those transporters or related proteins localized to the nucleus.
Although preliminary immunoblot analysis supports the immunofluorescent
nuclear localization of EAAT1, more extensive studies need to be performed. Various

51
membrane markers need to be used to determine whether or not the EAAT1
immunoreactivity is a result of the contamination of nuclear preparations with other
membranes. Future experiments could also include the isolation of the EAAT1
immunoreactive band from an SDS-PAGE gel for protein sequencing, as well as
immunofluorescence on isolated nuclear fractions.

Figure 3-1. Extracellular staining of human fibroblasts with EAAT3 antibody. Human
fibroblasts were fixed with 4% PFA and subjected to immunohistochemistry using the
procedures described in the Methods Chapter. Panel A demonstrates the clustering of
EAAT3 transporters on the surface of human fibroblasts using a polyclonal anti-EAAT3
antibody detected with a goat anti-rabbit IgG conjugated to FITC. Panel B shows cells
that were incubated with anti-EAAT3 antibody following preadsorption for 12 h at 4C
with 50 pg/ml of the corresponding peptide antigen. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
An outline was drawn around one cell to distinguish the periphery. The data shown
represent analysis of 0.2 pm sections through the cells.

53
10.0 um

Figure 3-2. Intracellular staining of human fibroblasts with EAAT3 antibody and co
localization with organelle-specific antibodies. Using the methods described in the
Methods Chapter, human fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with antibodies specific for EAAT3 (A), EAAT3 and KDEL (B),
or EAAT3 and transferrin receptor (TfR) (C). Co-localization of the proteins was
assayed by using simultaneously a rabbit polyclonal antibody against EAAT3 detected by
FITC-labeled goat anti-rabbit IgG and mouse monoclonal antibodies against KDEL and
TfR detected by Texas Red labeled goat anti-mouse IgG. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. The data shown represent analysis of 0.2 pm sections through the cells.

55

Figure 3-3. Nuclear staining of human fibroblasts with EAAT1-R antibody and co
localization with nucleus-specific antibodies. Using the methods described in the
Methods Chapter, human fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with antibodies specific for EAAT1-R (A), EAAT1-R that had
been preadsorption for 12 h at 4C with 50 pg/ml of the corresponding peptide antigen
(B), EAAT1-R and 414 (C), or EAAT1-R and D77 (D). Co-localization of the proteins
was assayed by using simultaneously a rabbit polyclonal antibody against EAAT1-R
detected by FITC-labeled goat anti-rabbit IgG, and mouse monoclonal antibodies against
414 and D77 detected by Texas Red labeled goat anti-mouse IgG. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. An outline was drawn around one cell in each panel to distinguish the
periphery. The data shown represent analysis of 0.2 pm sections through the cells.


Figure 3-4. Intracellular staining of human fibroblasts with EAAT1-S and EAAT1-C
antibodies. Using the methods described in the Methods Chapter, human fibroblasts were
fixed with -20C MeOH and subjected to immunohistochemistry with antibodies specific
for EAAT1-S (A) and EAAT1-C (B). The EAAT1-S antibody was detected by an FITC-
labeled goat anti-rabbit IgG, and the EAAT1-C antibody was detected by a Cy3-labeled
goat anti-guinea pig IgG. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells.

59

Figure 3-5. Intracellular staining of Hela cells with EAAT1-R and EAAT1-S antibodies
Using the methods described in the Methods Chapter, Hela cells were fixed with -20C
MeOH and subjected to immunohistochemistry with antibodies specific for EAAT1-R
(A) and EAAT1-S (B). The EAAT1-R and EAAT1-S antibodies were detected by an
FITC-labeled goat anti-rabbit IgG secondary antibody. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
Panel A shows the nuclear staining of multiple cells, whereas panel B is one cell
(outlined) with intracellular vesicle staining. The data shown represent analysis of 0.2
pm sections through the cells.

61

62
Q>
&
100,000g pellet
0)
o
CONFLUENCE
o
CO
50 75 100%
0

203
118
Figure 3-6. Immunoblot analysis of EAAT1 in the nuclear and intracellular membrane
fractions from human fibroblasts. A 30 ug aliquot of the 300 x g nuclear fraction (lane
1) and the 100,000 x g total intracellular membrane fraction (lanes 2-4) was subjected to
SDS-PAGE as described in the Methods Section of Chapter 3. Immunoblot analysis was
performed with a 1:1,000 dilution of EAAT1-R antibody and was detected with a
1:10,000 dilution of goat anti-rabbit IgG conjugated to horseradish peroxidase (HRP).
The 50, 75, and 100% labels indicates the degree of cell confluence at the time of cell
lysis and subfractionation. The blot shown is representative of three independent
experiments.

Figure 3-7. Expression of GFP and GFP-EAAT1 fusion proteins in human fibroblasts.
Human fibroblasts were transfected for 3 h with GFP(N3) only (A), or with the EAAT1-
GFP(N3) (B), and GFP(C3)-EAAT1 (C) fusion proteins according to the lipofectamine
protocol described in the Methods Chapter. Following 24-48 h of expression, cells were
fixed with -20C MeOH and visualized by deconvolution microscopy. Images were
processed from three independent transfections and the fluorescence patterns of expressed
proteins were determined to be reproducible. The data shown represent analysis of 0.2
pm sections through the cells.

64

Figure 3-8. EAAT1 immunofluorescent staining of human fibroblasts transfected with
EAAT1 -GFP(N3). Human fibroblasts were transfected for 3 h with the EAAT1 -
GFP(N3) fusion protein according to the lipofectamine protocol described in the Methods
Chapter. Following 24-48 h of expression, cells were fixed with -20C MeOH and
stained with antibodies against EAAT1-R (A) and EAAT1-S (B). Both EAAT1
antibodies were detected with a goat anti-rabbit IgG conjugated to Texas Red and
visualized by deconvolution microscopy. Images were processed from three independent
experiments and the staining was determined to be reproducible. The data shown
represent analysis of 0.2 pm sections through the cells.

66

Figure 3-9. EAAT1 immunofluorescent staining of PAEC transfected with EAAT1-
GFP(N3). PAEC were transfected for 3 h with the EAAT1-GFP(N3) fusion protein
according to the lipofectamine protocol described in the Methods Chapter. Following 24-
48 h of expression, cells were fixed with -20C MeOH and stained with antibodies
against EAAT1-R (A) and EAAT1-S (B). Both EAAT1 antibodies were detected with a
goat anti-rabbit IgG conjugated to Texas Red and visualized by deconvolution
microscopy. Images were processed from three independent experiments and the staining
was determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.

68

CHAPTER 4
LYSINURIC PROTEIN INTOLERANCE
Introduction
Lysinuric Protein Intolerance (LPI) is an autosomal recessive disease that is
characterized by a defect in dibasic amino acid transport, as well as, an impaired urea
cycle (reviewed by Simell, 1989). Although cases have been reported worldwide, the
highest prevalence is in Finland, where LPI afflicts 1 in 60,000 to 80.000 people.
Patients have an extremely low tolerance for dietary protein, and symptoms of
hyperammonemia are revealed shortly after weaning infants from the high-fat, low-
protein breast milk. Nausea, vomiting, and diarrhea following meals are early indications
of the disorder. Throughout infancy and childhood, patients show signs of growth
retardation and fail to thrive as a result of protein deficiency. They have enlarged livers
and spleens, muscle hypotonia and hypertrophy, osteoporosis, and approximately 20% of
the patients show varying degrees of mental retardation. The only available treatment is
to limit the consumption of protein. Normalization of hepatic nitrogen utilization and
urea synthesis requires supplementing meals with 3 to 8 grams of citrulline daily.
LPI was originally defined by elevated urinary levels and poor intestinal
absorption of all cationic and several neutral amino acids (Simell, 1989). Normal daily
urine contains a mean of 4.13 mmol lysine/1.73 m2 body surface area, whereas the urine
from LPI patients contains a mean of 25.7 mmol lysine/1.73 m2 body surface area
69

70
(Simell. 1989). The concentrations of cationic amino acids in plasma are low. whereas
glutamine, alanine, serine, proline, citrulline, and glycine are increased. Although
cationic amino acid transport is probably defective in most cell types, it has been
documented to be so in kidney tubules, intestine, and cultured fibroblasts (Simell, 1989).
LPI-derived fibroblasts accumulate elevated steady state levels of cationic amino acids
and exhibit a reduced rate of trans-membrane exchange for these same substrates (Smith
et al., 1987). The biochemical basis for this decreased release of cellular amino acids has
not been established. Although several amino acidurias are thought to result from
defective NaT-dependent transport across the brush border domain of either intestinal or
renal epithelial cells, previous studies have suggested that the primary transport defect in
LPI is a reduced efflux across the basolateral surface (Simell, 1989; Rajantie et ah. 1981).
Evidence for this interpretation comes from an observation that plasma amino acid
concentrations remain low following the oral administration of both cationic amino acids
and lysine dipeptides. Dipeptides are transported normally across the brush border
membrane by a mechanism distinct from that of free amino acids. The dipeptides are
then hydrolyzed to free amino acids by intracellular enzymes. However, in LPI
enterocytes, these free amino acids are unable to pass through the basolateral membrane
into the plasma, and instead, are transported back out via a brush border membrane
transporter. Direct measurements of lysine transport in intestinal biopsy specimens have
confirmed that the transport defect is located on the basolateral membrane (Simell, 1989).
Transport appears to be normal at the luminal surface of renal epithelial cells from LPI
patients (Simell, 1989).

71
Lvsine Transport in Whole Cells and Plasma Membrane Vesicles of LPI
System y" is a Na-independent activity that catalyzes facilitated transport and
exchange reactions (White. 1985), but also may permit accumulation of cationic amino
acids against a concentration gradient driven by the trans-membrane potential (Bussolati
et al., 1987). Under normal physiological conditions. System y~ mediates the uptake of
cationic amino acids such as arginine, lysine, and ornithine. CAT1 was the first cDNA to
be identified that exhibited System y* activity (Albritton et al., 1989). This transporter,
which also serves as the murine ecotropic leukemia virus receptor, mediates the Na -
independent, high-affinity transport of cationic amino acids in most mammalian cells
(Kim et al., 1991; Wang et al., 1991). The second member of the "CAT" family to be
isolated, CAT2. shares 61% identity with CAT1 at the amino acid level, and mediates
Na -independent, high-affinity transport of cationic amino acids in activated T
lymphocytes (MacLeod et al., 1990). The third clone to be identified, CAT2a, is the
result of alternative splicing of the CAT2 gene. It differs from CAT2 by only 41 amino
acids, yet it exhibits a 10-fold lower affinity for arginine and is liver-specific (Closs et al.,
1993). The last member of the "CAT" family, CAT3, shares the greatest homology with
CAT1 and mediates the Na-independent high-affinity transport of cationic amino acids
in the brain (Hosokawa et al., 1997; Ito and Groudine, 1997).
In cultured fibroblasts, there is little difference between the initial rates (measured
after 15 sec) of lysine uptake in normal and LPI cells, however, after 20 min, the LPI
cells accumulate two-fold more H-lysine than control cells (Handlogten and Kilberg,

72
unpublished results). The lack of a large difference in initial uptake rates suggests that
influx at the plasma membrane is unaffected by the disease. To eliminate the possible
confounding effects of trans-stimulation when measuring transport in whole cells, a crude
mixture of cellular membrane vesicles was isolated from cultured fibroblasts and used to
assay lysine uptake. Qualitatively, the data obtained from vesicle transport experiments
were similar to the results of whole cell transport. Once again, initial measurements
showed little difference in uptake by normal and LPI-derived vesicles, whereas LPI-
derived vesicles eventually accumulated four to five times more 'H-lysine than the
control vesicles (Handlogten and Kilberg, unpublished results). These results suggested a
decreased efflux of amino acid from either the plasma membrane or intracellular vesicles
of the LPI cells.
The observation that the LPI-derived vesicles also accumulate select neutral
amino acids (Handlogten and Kilberg, unpublished data) is consistent with the detection
of increased neutral amino acids in the urine of LPI patients. This prompted our
laboratory to investigate amino acid transport by other known transport systems in LPI
fibroblasts and plasma membrane vesicles. System b0 + is a NaMndependent system that
mediates the bidirectional transport of both cationic and neutral amino acids (Van
Winkle, 1988; Van Winkle et ah, 1988). However, this system is probably not
responsible for the elevated Na-independent uptake of neutral amino acids observed
during in vitro studies, because the transport of leucine was poorly inhibited by lysine in
the plasma membrane vesicles derived from LPI fibroblasts (Handlogten and Kilberg,
unpublished results). The NBAT protein, which exhibits the properties of System b0 +,

73
has been implicated as the defective cystine transporter in the inherited disease, cystinuria
(Calonge et ah. 1994).
System B0', first described by Van Winkle and colleagues (Van Winkle et ah.
1985), is responsible for cationic and neutral amino acid transport in various cell types,
including human fibroblasts. However, this is a Na*-dependent system, and therefore,
does not appear to be responsible for the increased LPI amino acid accumulation. System
y"L mediates the Na-independent high-affinity transport of cationic amino acids, as well
as the Na~-dependent high-affinity transport of L-leucine (Deves et ah, 1992). The
observation by our laboratory that 5 mM 2-amino-[2,2,l]-bicycloheptane-2-carboxylate
(BCH) inhibited leucine uptake in the normal and LPI vesicles is consistent with the
hypothesis that System y+L may be responsible for the unusual accumulation of cationic
and neutral amino acids in the LPI vesicles. L-leucine transport was inhibited by leucine,
lysine, arginine, and BCH in normal fibroblasts vesicles. However, lysine did not block
the L-leucine uptake in LPI vesicles. Collectively, the vesicle transport data is not
entirely consistent with the properties of any of the known transport systems. Therefore,
it is possible that an uncharacterized transport system is responsible for the activity
observed in the LPI vesicles. Alternatively, a known transport system that is altered in its
substrate specificity in the LPI cells may account for the elevated amino acid
accumulation.
Immunofluorescence Studies
When transport studies in LPI cells began, CAT1 was the only family member
that had been reported to mediate System y+ activity. Therefore, based on clinical

74
observations in LPI patients, it was originally hypothesized that LPI may arise from a
defect in the cationic amino acid transporter, CAT1, previously termed System y+ (Smith
et al., 1987). However, two independent laboratories sequenced the human CAT1
mRNA expressed in LPI patients and found no mutations within the sequence (personal
communication, Dr. Olli Simell, University of Turku). As a result, our laboratory
hypothesized that a trafficking defect involving the CAT1 transporter, or a protein
involved in membrane protein trafficking, may be responsible for the altered transport of
cationic amino acids in LPI cells. Preliminary immunofluorescence studies in our
laboratory revealed a population of intracellular vesicles, unique to LPI cells, which
appeared to contain the CAT1 transporter (Woodard and Kilberg, unpublished data). The
elevated steady-state accumulation of lysine, described above, supports the hypothesis
that amino acids may be sequestered in the abnormal intracellular vesicles of LPI cells. If
the defect involves a deficiency in efflux across the basolateral membrane domain in
epithelial cells, as proposed, perhaps the reason is that the lysine becomes trapped within
the intracellular vesicles instead of rapidly equilibrating across the plasma membrane.
This hypothesis is consistent with the finding by Smith et al. that trans-stimulation of
cationic amino acid efflux is also decreased in LPI cells (Smith et al., 1987).
Dr. Olli Simell, at the University of Turku (Turku, Finland), has been following
Finnish patients with LPI for over 25 years. His laboratory has thoroughly documented
the clinical aspects of the disease and is currently working to localize and characterize the
defective gene. Using linkage analysis and a candidate gene approach, Simell and
coworkers have excluded the possibility that CAT1 or CAT2 (Lauteala et al., 1997) is the
defective gene. Despite these observations, it is clear that the LPI fibroblasts exhibit

75
abnormal cationic and neutral amino acid transport and accumulate abnormal intracellular
vesicles and vacuoles. Therefore, this project has focused on the evaluation of membrane
protein trafficking in the cells of LPI patients.
Trafficking Defects in Plasma Membrane Transport Proteins
There is precedence for the aberrant trafficking of a specific class of transporter
proteins in both Saccharomyces cerevisiae and Drosophila. It was shown that S.
cerevisiae requires an endoplasmic reticulum (ER) integral membrane protein, SHR3, for
the effective processing and trafficking of amino acid permeases, specifically (Ljungdahl
et al., 1992). Mutations in SHR3 cause retention of 13 out of 13 amino acid permeases
tested within the ER, and therefore, block amino acid uptake by interfering with the
plasma membrane localization. The defective SHR3 does not affect the targeting of any
other plasma membrane, secretory, or vacuolar protein (Ljungdahl et. ah, 1992). In
Drosophila, mutations in a gene (ninaA) encoding an ER cis-trans isomerase cause the
abnormal intracellular accumulation of two homologous opsin proteins in photoreceptor
cells. The accumulation is presumed to occur as a result of improper protein folding early
in the biosynthetic pathway (Colley et. ah, 1991).
In humans, a single amino acid deletion in a highly regulated chloride channel is
the basis for the fatal genetic disorder cystic fibrosis (CF). It was shown that the
mutation that was present in 70% of the defective CF genes (Kerem et ah, 1989) resulted
in the abnormal retention of the cystic fibrosis transmembrane conductance regulator
protein (CFTR) in the endoplasmic reticulum, although the CF conductance activity
measured in ER vesicles or reconstituted proteoliposomes was unimpaired (Cheng et ah,

76
1990; Drumm et al.. 1991). Whereas both the wild-type and mutant CFTR proteins are
associated with calnexin in the ER, only the wild-type protein exits the ER and is
correctly trafficked to the plasma membrane (Pind et al., 1994).
Although no mutations have been found in the CAT1 transporter gene in LPI
patients, or any other protein at this point, the LPI defect results in pleiotropic effects on
amino acid transport (Simell, 1989). Broad scope changes in transport could result from
a block of transporter processing at any stage along the various pathways of membrane
protein synthesis or endocytic recycling. This chapter describes the cellular localization
of the CAT1 arginine transporter and of various organelle-specific proteins along the
biosynthetic, endocytic. and degradative pathways in normal and LPI fibroblasts.
Results
Morphology of normal and LPI fibroblasts by light and electron microscopy.
There are basic differences between the normal and LPI cells that can be discerned at the
light microscope level (Figure 4-1). The LPI cells frequently appear to be larger with
long processes and extensions at the cell periphery. Also, normal fibroblasts grow
significantly faster than their LPI counterparts under the same culture conditions until
they reach confluency. The LPI cells tend to grow in clusters and never reach full
confluence. The most striking difference between the normal and LPI cells is a
population of large vesicles or vacuoles observed throughout the cytoplasm, often
concentrated around the nucleus of the LPI cells. When electron microscopy was used to
increase the resolution of these large intracellular vacuoles, at a magnification of 25K
they appear to contain a fibrous material of unknown origin (Figure 4-2). The unusual

77
LPI-derived vacuoles have been detected by electron microscopy in the cell lines from
two different LPI patients (Woodard and Kilberg. unpublished data; McDonald and
Kilberg, unpublished data). The only previous microscopic observation reported was in a
paper by Simell and coworkers who mentioned apparent alterations in the architecture of
hepatocytes (Simell et al., 1975). They reported an accumulation of vesicles containing a
"fibrillogranular material." although it is not known whether these represent the same
vesicles we have identified.
CAT1 antibody production against human. The preliminary immunofluorescence
research was performed using a CAT1 polyclonal antibody that was generated against a
25 amino acid sequence (SIKNWQLTEKNFSCNNNDTNVKYGE) from the third
extracellular loop of the murine CAT1 sequence (Woodard et al., 1994). The murine
CAT1 antibody was shown to stain specifically a wide variety of cell types from several
species and was inhibited by pre-incubation with the corresponding peptide. However,
this anti-murine CAT1 antibody did not immunoblot well, and it is possible that it is not
optimal for detection of denatured protein in the human cell lines being tested. For this
reason, a polyclonal antibody was generated against a sequence of 20 amino acids
(CEEASLDADQARTPDGNLDQ) at an intracellular site of the human CAT1 protein
(Cocalico Biologicals, Inc.. Reamstown, PA). Enzyme-linked immunosorbent assays
(ELISA) were performed on the individual serum samples to determine antibody titers.
Briefly. 96-well plates were coated with the human CAT1 peptide (above), and incubated
with a 1:50 to a 1:32,000 dilution of the immune or pre-immune serum collected from the
inoculated rabbit. Following the removal of the primary antibody, an alkaline-
phosphatase conjugated secondary antibody was added to the plates and detected with a

78
phosphatase substrate at a wavelength of 405 nm. By the sixth bleed, one of the human
CAT1 antibodies showed a 2.5-fold increase in absorbance over the pre-immune at a
dilution of 1:50. Although a series of immunohistochemistry experiments were
conducted using this anti-human CAT1 antibody, the background staining was high and
the corresponding peptide failed to completely inhibit staining. Therefore, all of the data
presented in this chapter was generated using the polyclonal murine CAT1 antibody. The
conditions for the CAT1 antibody as well as the other antibodies used in this chapter are
summarized in Table 4-1.
Table 4-1
Antibodies for Immunofluorescence Studies
Name Host Source Dilution
CAT1
rabbit
EAAT1
rabbit
EAAT3
rabbit
GLUT1
rabbit
414
mouse
D77
mouse
P-integrin
mouse
caveolin
rabbit
tubulin
mouse
KDEL
mouse
BiP
rabbit
PDI
rabbit
sialyltransferase
rabbit
Dr. Michael Kilberg,
Univ. of Florida
1:25
Dr. Jeffrey Rothstein.
John Hopkins
1:50
Dr. Michael Kilberg.
Univ. of Florida
1:200
Dr. Susan Frost.
Univ. of Florida
1:50
Dr. John Aris,
Univ. of Florida
1:10
Dr. John Aris.
Univ. of Florida
1:50
Dr. Martin Hemler,
Dana-Farber Cancer Inst.
1:100
Transduction Laboratories
Lexington, K.Y
1:200
Sigma.
St. Louis, MO
1:100
Dr. David Vaux.
EMBL
1:100
Dr. Susan Frost.
Univ. of Florida
1:50
Dr. Tom Wileman,
Dana-Farber Cancer Inst.
1:25
Dr. William Dunn.
Univ. of Florida
1:50

79
Name
Host
Source
Dilution
M6P receptor
rabbit
Dr. Peter Nissley,
NIH
1:50
Rab 5
rabbit
Santa Cruz Biotechnology
Santa Cruz, CA
1:100
transferrin receptor
mouse
Zymed.
San Francisco, CA
1:5
lpgl20
Rabbit
Dr. William Dunn,
Univ. of Florida
1:100
cathepsin D
Rabbit
Biodesign International,
Kennebunk, ME
1:300
Distribution of endogenous CAT1 in normal and LPI fibroblasts. As mentioned
above, the LPI disease was originally characterized by elevated concentrations of arginine
and lysine in the urine. Later, it was determined that the transport defect is expressed in
the kidney tubules, intestines, cultured fibroblasts, and probably hepatocytes (Simell,
1989). It has been shown that CAT2 and CAT2a are not expressed in human fibroblasts,
so our laboratory began to investigate the localization of the CAT1 transporter in normal
and diseased cells. Normal and LPI fibroblasts were fixed with 4% PFA and labeled with
a 1:25 dilution of the CAT1 transporter antibody (according to the protocol in the
Methods chapter). Antibody staining was detected with a 1:200 dilution of goat anti
rabbit IgG linked to FITC. Both normal and LPI fibroblasts demonstrated an
extracellular periodic labeling that resembled intensely stained patches on the plasma
membrane (data not shown). Staining of the fluorescent patches was completely blocked
when the murine CAT1 antibody was pre-incubated with 50 pg/ml of corresponding
peptide for 12 hours before labeling. This same pattern was observed in several different
LPI fibroblast cell lines, as well as, in porcine pulmonary artery endothelial cells (see
Chapter 5). Incubation of the cells with the microtubule inhibitor, nocodazole, caused the

80
transporter staining to disperse over the entire cell surface, and removal of the inhibitor
caused the clusters of transporter to reform within 3 h (data discussed in Chapter 5).
These data confirmed previous reports by Woodard et al. that the CAT1 antibody forms
clusters on the plasma membrane of normal fibroblasts, and that this arrangement is
dependent on intact microtubules (Woodard et al., 1994). Identification of these clusters
as caveolae will be discussed in Chapter 5.
In a separate series of experiments, the intracellular distribution of the CAT1
transporter was detected by immunofluorescence and deconvolution microscopy in
MeOH-fixed fibroblasts. Although an intracellular vesicle population was labeled with a
1:25 dilution of the CAT1 antibody in both normal (Figure 4-3 A) and LPI fibroblasts
(Figure 4-3 B), there was not a significant difference in the number or physical
appearance of the vesicles. Incubating cells with secondary antibody only, or incubating
the primary antibody with 50 pg/ml of corresponding peptide, resulted in only a faint,
diffuse background immunofluorescence (data not shown). These results are different
than earlier reports from our laboratory that documented an obvious difference between
the normal and LPI cells when observed using an epifluorescence microscope. It was
observed previously that CAT1 antibody labeling of normal fibroblasts resulted in diffuse
cytoplasmic staining with slightly increased intensity in the region of the Golgi complex
(a perinuclear pattern). On the other hand, the CAT1 antibody staining in the LPI cells
was specific for a large number of intracellular vesicles, which appeared to be randomly
distributed throughout the cytoplasm. As a result of the apparent discrepancy, staining of
normal and LPI fibroblasts was repeated using a variety of conditions. CAT1 serum.

81
IgG-purified, and affinity-purified antibodies were used at dilutions ranging from 1:25 to
1:1000. In addition, cells were fixed with either -20C MeOH or 2%-4%
paraformaldehyde, for 10 to 30 minutes, before permeabilizing and incubating with the
CAT1 antibody. Several different normal and LPI patient cell lines were used and
antibody incubations ranged from 1 h to overnight (data not shown). However, none of
these conditions produced the CATI-containing LPI vesicles seen previously. Additional
research, perhaps with new antibody preparations, will be required to confirm the CAT1
staining of LPI fibroblasts.
Previous LPI research in the Kilberg laboratory was limited to the study of the
CAT1 transporter, yet it is entirely possible that the proposed trafficking defect involves
other transporter, and/or non-transporter, proteins. Antibodies against several other
nutrient transporters were available and used for intra- and extracellular
immunofluorescent labeling of normal and LPI fibroblasts. Both cell types were stained
with a 1:50 dilution of the NaMndependent glucose transporter, GLUT1, a 1:200 dilution
of EAAT3, and a 1:50 dilution of EAAT1 following both -20C MeOH and 4%
paraformaldehyde fixation. Each of these antibodies was detected using a goat anti-rabbit
IgG conjugated to FITC. GLUT1 labeling was detected in small amounts on the plasma
membrane, but primarily in cytoplasmic and perinuclear vesicles. The EAAT3 antibody
was detected in clusters on the plasma membrane, as well as in vesicles throughout the
cytoplasm. The EAAT1 antibody, discussed in detail in Chapter 3, was localized to the
nuclear membrane. No differences in the staining patterns were detected, and no
abnormal intracellular vesicle populations were identified, when labeling the normal and
LPI cells with antibodies against any of these transporters (data not shown).

82
Localization of expressed CAT1 in normal and LPI fibroblasts. The experiments
investigating endogenous CAT1 localization relied on the detection of the human
transporter with a anti-murine CAT1 antibody. An expression system was developed in
order to confirm endogenous experiments, as well as, to avoid any technical problems
associated with the CAT1 antibody. Transfection with several CAT1 constructs was
attempted before deciding to use the green fluorescent protein (GFP) vector from
Clontech (Palo Alto, CA). These constructs included a FLAG-tagged murine CAT1
cDNA, generated in our laboratory, as well as an HA-tagged human CAT1 cDNA and a
human CAT1-GFP fusion protein, both provided by Dr. Lorraine Albritton at University
of Tennessee Medical Center. Neither the HA-tagged CAT1 nor the FLAG-tagged CAT1
constructs were detected following transfection, and the expression of the CAT 1-GFP
from Dr. Lorraine Albritton was extremely low (data not shown). Dr. Lorraine Albritton's
CAT 1-GFP fusion protein was constructed with a GFP variant that was not modified for
enhanced expression and fluorescence in human cells. Therefore, a GFP(C3)-CAT1
fusion protein was constructed using Clontechs humanized GFP variant and the CAT1
cDNA sequence from mouse (See Methods Chapter).
The GFP tag (described in the Methods Section) provided an alternative technique
for observing the localization of the CAT1 protein in the normal and LPI fibroblasts. If
the CAT1 transporters were trafficking incorrectly in LPI cells, but the available antibody
could not detect it, then the GFP(C3)-CAT1 would provide an independent method of
CAT1 localization. Transfection of normal fibroblasts with the GFP(C3) only, which
lacks a targeting signal and, therefore, diffuses throughout the cytoplasm and nucleus,
was used as a control (Figure 4-4 A). Alternatively, transfection of normal fibroblasts

83
with the GFP(C3)-CAT1 demonstrated specific targeting to the plasma membrane, as
well as one or more intracellular vesicle populations (Figure 4-4 B). The strong
perinuclear staining is probably GFP(C3)-CAT1 in the Golgi. The expressed GFP(C3)-
CAT1 fusion protein also showed significant co-localization with the CAT1 antibody
(detected with a 1:200 dilution of goat anti-rabbit IgG linked to Texas Red) in normal
fibroblasts (Figure 4-4 C). Although the GFP(C3)-CAT1 made it easier to visualize the
exact location of the CAT1 transporter, there was no difference in the intracellular pools
of GFP-CAT1 in normal (Figure 4-5 A) and LPI (Figure 4-5 B) fibroblasts. In both cell
lines, the GFP-CAT1 fusion protein was detected on the plasma membrane, throughout
the cytoplasm in small vesicles, and highly concentrated in the perinuclear region
(probably representing Golgi).
Examination of organelle integrity in normal and LPI fibroblasts. As discussed,
no significant differences in the distributions of any of the amino acid transporters were
identified in the normal and LPI fibroblasts. However, given the differences in
morphology observed between normal and LPI fibroblasts and the presence of the LPI-
specific vacuoles, it is clear that the disorder is associated with a basic cellular defect that
can be visualized as a structural deformity in one or more of the organelles of the LPI
cells. Therefore, staining patterns generated by organelle-specific antibodies, in normal
and LPI cells, were compared in order to gain important information regarding the
integrity of the organelles in the LPI fibroblasts. A variety of organelle-specific
antibodies were used as markers for the identification of compartments involved in the
biosynthetic, endocytic, and degradative pathways. The antibodies that were chosen are
prototypical markers for membrane protein trafficking compartments and have been

84
described in the literature for a variety of cell lines. Initial experiments were performed
in order to confirm the reliability of the antibodies and document the staining patterns in
the normal human fibroblasts. For all of the immunoassays, secondary antibodies used
were either goat anti-rabbit IgG or goat anti-mouse IgG conjugated to either fluorescein
isothiocyanate (FITC, Sigma Chemical Co., St. Louis, MO) or Texas Red (TR, Cappel
Laboratories, Durham, NC).
The nuclear membranes of normal (Figure 4-6 A) and LPI (Figure 4-6 C) cells
were examined using a 1:10 dilution of mouse anti-human 414 antibody, which was
generated against an epitope common to several of the nuclear pore complex proteins
(Davis and Blobel, 1986). A 1:50 dilution of mouse anti-yeast D77 antibody (Aris and
Blobel, 1988) was used to visualize the nucleoli of both normal (Figure 4-6 B) and LPI
(Figure 4-6 D) fibroblasts. The labeling of the nuclear structures was consistent with
images from the literature, and there was no discernible difference between the staining
patterns of the normal and LPI fibroblasts. A 1:100 dilution of mouse anti-human-p-
integrin antibody was used to label the plasma membrane (Figure 4-7 A and B), and a
1:200 dilution of rabbit anti-caveolin-1 antibody was used to specifically detect the
caveolar domains of the plasma membrane (Figure 4-7 C and D). In both normal and LPI
cells, the caveolin antibody labeled specific regions of the plasma membrane, as
expected, and the P-integrin antibody showed strong staining around the periphery of
both cell types. Although the LPI cells may be defective in cell-to-cell contact, indicated
by the way they grow in clusters rather than spreading out to confluence, there was no
noticeable difference in the P-integrin staining pattern of the normal and LPI cells. The

85
microtubules appeared to be intact in both the normal (Figure 4-7 E) and LPI cells
(Figure 4-7 F) according to immunofluorescent labeling using a 1:100 dilution of mouse
anti-P-tubulin antibody.
Antibodies against resident proteins of the endoplasmic reticulum (ER) and Golgi
were used to compare the organelles of the biosynthesis pathway in normal and LPI cells.
A short, four amino acid (Lys-Asp-Glu-Leu) peptide, called KDEL, appears in the
sequences of ER resident proteins (such as BiP) and is responsible for selectively
retrieving the proteins after they leave the ER in transport vesicles (reviewed by Pelham,
1991). A membrane-bound receptor in the cis-Golgi recognizes the KDEL retention
signal, and returns the KDEL-containing proteins to the ER. The ER, in normal and LPI
fibroblasts, was detected using a 1:100 dilution of mouse anti-KDEL antibody (Figure 4-
8 A and B), a 1:50 dilution of rabbit anti-BiP antibody (data not shown), as well as, a
1:25 dilution of rabbit anti-protein disulfide isomerase (PDI) antibody (data not shown).
The KDEL antibody provided the best labeling, but each of the antibodies stained the ER
with no significant difference between the two cell lines. The ER in the LPI cells
appeared to be larger and more spread out than the ER of the normal fibroblasts;
however, this was probably due to the fact that the LPI cells tend to be larger in general.
Compartments of the Golgi were stained using 1:50 dilutions of either rabbit anti-
sialyltransferase antibody (Figure 4-8 C and D), which labels the Golgi and Trans-Golgi
Network (TGN), or rabbit anti-mannose-6-phosphate receptor antibody (Figure 4-8 E and
F), which is specific for TGN and late endosomes. There was no recognizable difference

86
in the structure of the Golgi, or any component of the biosynthetic pathway that was
investigated in the normal and LPI fibroblasts.
Organelles of the recycling pathway were visualized using a 1:100 dilution of
rabbit anti-human Rab5 antibody (Figure 4-9 A and B), for labeling early endosomes, and
a 1:5 dilution of mouse anti-human transferrin receptor antibody (Figure 4-9 C and D),
for detecting vesicles involved in endocytosis and recycling. Neither of these antibodies
detected an abnormality in the recycling pathway of the LPI cells. The number of
vesicles containing the transferrin receptor varied significantly between experiments, but
the variation was independent of normal versus LPI. Vesicles and compartments
involved in degradation were identified using a 1:300 dilution of the lysosomal protease,
cathepsin D (Figure 4-10 A and B). Although the antibody demonstrated a punctate
staining pattern in both cell types, the vesicles detected in the LPI fibroblasts (Figure 4-10
B) were larger and in greater abundance than those observed in the normal fibroblasts
(Figure 4-10 A). To further test for a difference in the lysosomal staining, normal and
LPI cells were labeled with a 1:100 dilution of a mouse antibody generated against the
lysosomal membrane protein, lpgl 20 (Figure 4-11 A and B). In the normal fibroblasts
(Figure 4-11 A), a punctate pattern of lysosomal labeling was observed throughout the
cytoplasm. On the other hand, the apparent diameter of the lysosomes detected in the LPI
fibroblasts (Figure 4-11 B) were larger and in greater abundance than in the normal cells.
In addition, the LPI lysosomes were tightly clustered and located in close proximity to the
nucleus.
Treatment of normal and LPI fibroblasts with lysosomotropic agents. From the
labeling, it appeared as though the lpg!20-containing lysosomes corresponded to the

87
large "vacuoles" that were previously observed by phase contrast and electron
microscopy. Acridine orange (AO) is a lysosomotropic weak base that accumulates in
acidic compartments of living cells (Robbins et al 1963). When excited with blue light,
it emits a red fluorescence that can be visualized using a Texas Red filter. To determine
whether or not the lpgl20-containing compartment was acidic in nature, as would be
expected for lysosomes, normal and LPI fibroblasts were loaded with 5 pg/ml acridine
orange for 15 min before fixing with either 4% paraformaldehyde or -20C MeOH and
viewing with a Nikon axiophot inverted epifluorescent microscope (Figure 4-12 A and
B). Deconvolution microscopy could not be performed because the acridine orange
diffused out of the acidic compartments too rapidly. For the same reason, staining with
the lpgl20 antibody after loading with acridine orange was unsuccessful. However, it did
appear as though the AO-containing compartment in the LPI cells was larger and more
abundant (Figure 4-12 B), even though it was not confirmed that these structures
represented the lpgl20-containing lysosomes. The same AO staining patterns were
obtained when live normal or LPI cells were loaded with AO and immediately viewed
with the epifluorescence microscope. When exposed to an extremely high dose of AO
(ranging from 50-500 pg/ml), the normal fibroblasts died immediately, as judged by the
way they curled and lifted off the tray, whereas the LPI cells were able to tolerate the
drug (data not shown).
The lysosomes of normal and LPI fibroblasts were also compared following
treatment with chloroquine, a lysosomotropic drug that neutralizes lysosomes and other
acidic compartments. Both cell lines were incubated with 50 pM of chloroquine in MEM

88
+ 10% FBS for 1 h at 37C before fixing with -20C MeOH and staining with anti-lpgl20
antibody (Figure 4-13 A-D). The normal fibroblasts, without chloroquine treatment,
contained small lysosomes distributed throughout the cytosol, as visualized by labeling
with the lpgl 20 antibody (Figure 4-13 A). However, following chloroquine treatment,
the lpgl20-containing compartment in the normal fibroblasts (Figure 4-13 B) increased in
size and number resembling the lysosomal-like vacuoles of the untreated LPI cells
(Figure 4-13 C). When the LPI fibroblasts were treated with chloroquine, the lpg 120-
containing vacuoles grew to an enormous size and clustered tightly in the perinuclear
region of the cells (Figure 4-13 D). Incubating both cell types with a higher
concentration of chloroquine, or for an extended period of time resulted in similar, yet
even more pronounced effects. The vacuoles in the normal fibroblasts became even
larger and more abundant, whereas the intracellular membranes of the LPI cells appeared
to dissolve until only a few giant vacuoles filled the entire intracellular space (data not
shown).
Antibody labeling of normal and LPI fibroblasts expressing GFP-CAT1. The
experiments discussed in this section are an extension of those presented above. As
mentioned before, no differences were detected in the abundance or distribution of
endogenous or exogenous CAT1 in normal and LPI fibroblasts. Of all of the organelle-
specific antibodies tested, only those associated with the lysosomes showed any
abnormalities in the LPI cells. The following co-localization experiments examined
whether or not the CAT1 transporter, in LPI cells, was trapped in a particular
compartment or vesicle pool, such as the lysosomes. Normal and LPI fibroblasts were
transfected with the GFP-CAT1 fusion protein according to the lipofectamine protocol

89
described in the Methods Chapter. After 24 h, the cells were fixed with -20C MeOH,
and stained with organelle-specific antibodies (according to normal transfection and
immunofluorescence procedures described in the Methods Chapter). The GFP-CAT1
transfection was used instead of the CAT1 antibody because many of the organelle-
specific antibodies were generated in the same species as the CAT1 antibody, and
therefore, could not be used for double-labeling experiments. The Rab5 antibody was
used to co-localize GFP-CAT1 with early endosomes, the mannose-6-phosphate receptor
antibody was used to detect GFP-CAT1 in Golgi and late endosomes/TGN, and the
lpgl 20 antibody was used to observe GFP-CAT1 associated with the late endosomes and
lysosomes. The results obtained were the same for experiments conducted in both normal
and LPI fibroblasts. The GFP-CAT1 co-localized to a moderate degree with Rab5 and
mannose-6-phosphate receptor antibodies (data not shown); however, very little overlap
was observed with the lpgl20 antibody (Figure 4-14 A and B). This was a reasonable
result given the probable participation of the transporter in the biosynthetic and recycling
pathways. Even though the lysosomes, or a lysosome-like vesicle population, appear to
be abnormal in the LPI cells, there was no accumulation of the GFP-CAT1 in these
compartments.
Discussion
Previous experiments demonstrated that LPI fibroblasts exhibited an elevated
accumulation of cationic amino acids over time, as well as the existence of a unique
vesicle population. As a result, it was hypothesized that the CAT1 transporter may be
trapped in an intracellular compartment and, therefore, unable to mediate efflux of

90
cationic amino acids across the basolateral plasma membranes of affected cells. This
transporter defect would provide an explanation for the increased intracellular
accumulation of cationic amino acids during LPI whole cell and vesicle transport assays.
Experiments were performed to test for the presence of CAT1, or other transporters, in
abnormal LPI vesicles using a higher-resolution microscope than had previously been
available. In addition, organelle-specific antibodies were used to compare the cellular
compartments involved in biosynthesis, endocytosis, and degradation in normal and LPI
cells. Immunofluorescence studies did not detect an association of CAT1 transporters
with the abnormal LPI-associated vesicle population previously reported. In addition, the
extracellular and intracellular distributions of the CAT1, EAAT1, EAAT3, and GLUT1
transporters, as well as the integrity of specific organelles involved in the trafficking
pathways of membrane proteins appeared normal in the LPI fibroblasts. Therefore, the
most important difference between the normal and LPI fibroblasts, detected by
immunofluorescence in this study, was an abnormal population of enlarged lysosomal-
like vacuoles (described above).
It is confusing as to why earlier immunofluorescence data documenting the
CAT 1-containing LPI vesicles could not be reproduced. However, several experimental
conditions were changed that may explain the apparent discrepancy. First, although the
primary cell line used in this study came from the same patient as before, the cells were
isolated and cultured at different times. Although it is unlikely, the presence of CAT1 in
the LPI vesicles could have involved a defect specific to the cell line used in the first set
of experiments. Second, the same sample of CAT1 antibody was not available for this
study. The CAT1 peptide and antibody were older during the experiments described

91
above and may not have been as effective. As a result, new peptide should be prepared
and affinity-purified antibody generated for future experiments. Third, previous
immunofluorescence experiments were visualized using an epifluorescence microscope,
whereas the image processing for this project involved the use of a deconvolution
microscopy system. The ability to scan the cell in the z-plane, and the reduced
background provided by the deconvolution microscope, results in a higher degree of
resolution for antibody labeling. Dissection of the LPI cells in the z-plane may be
particularly important because of their tendency to shed pieces of the plasma membrane
and collect debris on their cell surface. Our laboratory has shown that this debris contains
CAT1 transport activity, stains with CAT1 antibody and thus, may create a problem for
detection of "intracellular" CAT1 when a standard epifluorescence microscope is being
used. The deconvolution microscope allows the user to non-invasively slice the cell into
sections from top to bottom. One or more sections can be taken from the middle of the
cell, thus avoiding fluorescent interference by cell surface debris.
Electron microscopy confirmed the presence of the large vacuoles previously
observed with light microscopy in the cytoplasm of LPI cells. Immunofluorescence
results indicate that the large vacuoles are probably lysosomes, or a related compartment
that contains cathepsin D and the lpgl 20 lysosomal membrane protein. When cells were
incubated with chloroquine, a reagent that neutralizes acidic compartments, the disease
phenotype of enlarged cytoplasmic vacuoles was induced in normal fibroblasts, and
exaggerated in LPI cells. Although lpgl 20 antibody staining of acridine orange loaded
cells was not possible, due to the rapid diffusion of the lysosomotropic drug, the pattern
of AO fluorescence indicated that the abnormal LPI lysosomes/vacuoles are probably

92
acidic in nature. The above observations all suggest that abnormal lysosomal-like
structures may be one of the problems contributing to the LPI disorder.
Interestingly, the LPI fibroblasts exhibit several morphological characteristics that
appear in the fibroblasts of patients with Chediak-Higashi Syndrome (CHS). CHS is an
autosomal recessive disorder that is characterized by clusters of giant lysosomes
concentrated in the perinuclear region of several cell types, including fibroblasts (Jones et
al., 1992). The gene responsible for this disorder has been cloned, yet little is known
about the function of the protein it encodes. The Beige protein (the mouse homologue to
the CHS gene) is believed to be a 400 kDa cytosolic protein that is expressed in most
mouse tissues (Perou et al., 1997). A deficiency in the Beige protein leads to the
formation of abnormally large lysosomes, whereas overexpression of the protein results
in unusually small lysosomes. This finding, along with results from a lysosome-
lysosome fusion assay (Ward et al., 1997) suggests that the CHS/Beige protein plays a
role in the fission (and not fusion) of lysosomal vesicles. However, there is no evidence
that the CHS/Beige protein interacts with lysosomes, or any other organellar membranes
(Perou et al., 1997). The CHS/Beige protein contains no obvious signal sequences,
transmembrane spanning domains, or membrane-anchors. The presence of WD40 repeats
(a known protein interaction domain) within the CHS/Beige protein sequence suggests
that it may interact with other proteins. Therefore, the CHS/Beige protein may
participate in lysosomal fission indirectly, via the interaction of another protein.
No CHS/Beige-associated proteins have been identified, however, preliminary
experiments by Davies et al. demonstrate that the down-regulation of either Rab7 or Rab9
proteins induces fibroblasts to form large vacuoles resembling those of the CHS

93
fibroblasts (Davies et al., 1997). Rab7 resides in late endosomes and is believed to
participate in the trafficking of proteins from early to late endosomes. Rab9 has been
implicated in the recycling of the mannose-6-phosphate receptor from late endosomes to
the TGN. Because of their involvement in the recycling and degradative pathways, as
well as, the morphological changes in lysosomes upon their down-regulation, Rab7 and
Rab9 could be candidates for association with the CHS/Beige protein. CHS patients
suffer from severe immunologic defects, abnormal platelet function, and partial ocular
and cutaneous albinism (Perou et al., 1997). Although the clinical symptoms of LPI are
different from those of CHS, the perinuclear accumulation of abnormally large vesicles in
the LPI fibroblasts is remarkably similar to the abnormal lysosomal vacuoles of CHS
fibroblasts.
There are several lysosomal storage disorders that result from the abnormal
accumulation of small molecules in the lysosomes (reviewed by Chou et al., 1992). The
sialic acid storage diseases represent a collection of inherited disorders characterized by
the elevated excretion of sialic acid in the urine, as well as, the accumulation of the acidic
monosaccharides in lysosomes (Mancini et al., 1991). Mancini and coworkers identified
and characterized a proton-driven sialic acid carrier with broad specificity to other acidic
monosaccharides in the lysosomal membranes of rat liver (Mancini et al., 1991).
Therefore, a defect in this carrier may represent the molecular basis underlying one or
more of the sialic acid storage diseases.
A number of amino acid transport systems have also been characterized in the
lysosomal membranes of human fibroblasts and rat liver (Chou et al., 1992). System c
represents a carrier-mediated cationic amino acid lysosomal transporter that shares many

94
of the characteristics with the plasma membrane System y+. System c mediates the Na-
independent transport of cationic amino acids and exhibits the property of trans
stimulation. Cystinosis is an autosomal recessive disease that results from defective
efflux of cystine from the lysosomes (Schneider and Schulman, 1982). Administering the
therapeutic agent, cysteamine, is the primary mode of treatment for this disorder. This
therapy is presumed to be successful because System c is capable of lowering lysosomal
cystine concentrations by allowing efflux of the mixed disulfides, cysteine and
cysteamine, out of the lysosomes (Pisoni et al., 1985). It is possible that lysine or other
positively charged amino acids are accumulating in the lysosomes of LPI cells due to a
defect in the lysosomal System c cationic amino acid transporter. Abnormal
accumulation of amino acids could be responsible for the swelling of the lysosomes in the
LPI fibroblasts. It has been reported that the incubation of lysosomes with chloroquine
leads to a reduction in lysine efflux. This decreased amino acid efflux may explain why
incubating the normal and LPI fibroblasts with chloroquine resulted in the swelling of the
lysosomal-like vesicles. If the LPI lysosomes have an elevated amount of lysine due to a
partially defective cationic amino acid transporter, then blocking the efflux further, with
chloroquine, may lead to an exaggerated phenotype of the disease. In addition, blocking
the functional lysosomal transporter in the normal fibroblasts with chloroquine treatment
would induce the disease phenotype.
One of the future directions of this project may be to prove whether or not LPI is a
lysosomal disease. Lysosomes isolated from normal and LPI cells, by differential
centrifugation, can be measured by flow cytometry (Perou et al., 1997). In addition, the
lysosomal-enriched preparation can be used for measuring the influx and efflux of

95
cationic and neutral amino acids. This may indicate whether or not specific amino acids
are accumulating in, and leading to the enlargement of, lysosomes in the LPI cells.

Figure 4-1. Morphology of normal and LPI fibroblasts by light microscopy. Normal (A)
and LPI (B) fibroblasts were grown under normal culture conditions according to the
protocol in the Methods Chapter and visualized in the 75-mm culture tray using a Nikon
Axiophot epifluorescence inverted microscope. The digitized image was captured using a
Spot CCD camera with a resolution of 1315 x 1033 pixels (Diagnostics Instruments, Inc.,
Sterling Heights, MI). Fluorescence from three independent experiments was analyzed
and shown to be reproducible.

97

Figure 4-2. Morphology of normal and LPI fibroblasts by electron microscopy. Normal
and LPI fibroblasts were fixed in a solution of 4% paraformaldehyde and 0.5%
glutaraldehyde for 30 min before pelleting. The cells were then dehydrated in ethanol,
embedded in an acrylic resin (either K4M or unicryl), and thin-sectioned for mounting on
nickel grids according to the protocol of the Electron Microscopy Core (University of
Florida). Electron micrographs were taken of normal fibroblasts at a magnification of 6K
(A), LPI fibroblasts at 6K (B), LPI fibroblasts at 9K (C), and LPI fibroblasts at 25K (D).

99

Figure 4-3. Intracellular staining of normal and LPI human fibroblasts with the CAT1
antibody. According to the procedures described in the Methods Chapter, normal (A) and
LPI (B) fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with an antibody specific for CAT1. The CAT1 antibody was
visualized by an FITC-labeled goat anti-rabbit IgG secondary antibody. Staining from
three independent experiments was analyzed by deconvolution microscopy and shown to
be reproducible. An outline was drawn around one cell to distinguish the periphery. The
data shown represent analysis of 0.2 pm sections through the cells.

101
10.0 um

Figure 4-4. Expression of GFP and the GFP(C3)-CAT1 fusion protein in normal human
fibroblasts. Normal fibroblasts were transfected for 3 h with GFP(C3) only (A), or with
the GFP(C3)-CAT1 fusion protein (B) according to the lipofectamine protocol described
in the Methods Chapter. Following 24-48 h of expression, cells were fixed with -20C
MeOH and visualized by deconvolution microscopy. In Panel C, GFP(C3)-CAT1
transfected cells were stained with an antibody against CAT1 and detected with a Texas
Red-labeled goat anti-rabbit IgG. Images were processed from three independent
transfections and the fluorescence patterns of expressed proteins were determined to be
reproducible. The data shown represent analysis of 0.2 pm sections through the cells.

eoi

Figure 4-5. Expression of the GFP(C3)-CAT1 fusion protein in normal and LPI human
fibroblasts. Normal (A) and LPI (B) fibroblasts were transfected for 3 h with the
GFP(C3)-CAT1 fusion protein according to the lipofectamine protocol described in the
Methods Chapter. Following 24-48 h of expression, cells were fixed with -20C MeOH
and visualized by deconvolution microscopy. Images were processed from three
independent transfections and the fluorescence patterns of expressed proteins were
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.

105

Figure 4-6. Intracellular staining of normal and LPI fibroblasts with antibodies against
proteins of the nuclear membrane and nucleolus. According to the protocol described in
the Methods Chapter, normal (A and B) and LPI (C and D) fibroblasts were fixed with
-20C MeOH and subjected to immunohistochemistry with the anti-414 antibody specific
for the nuclear membrane (A and C) and the anti-D77 antibody specific for the nucleolus
(B and D). Both primary antibodies were visualized with a goat anti-mouse IgG
conjugated to FITC. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. One cell in each panel is
outlined to distinguish the cell periphery. The data shown represent analysis of 0.2 pm
sections through the cells.

107

Figure 4-7. Intracellular staining of normal and LPI fibroblasts with antibodies against
plasma membrane and cytoskeletal proteins. According to the protocol described in the
Methods Chapter, normal (A, C, and E) and LPI (B, D, and F) fibroblasts were fixed with
-20C MeOH and subjected to immunohistochemistry with the anti-P-integrin antibody
specific for the plasma membrane (A and B), the anti-caveolin antibody specific for
caveolae (C and D), and the anti-p-tubulin antibody specific for microtubules (E and F).
The anti-P-integrin and anti-P-tubulin primary antibodies were visualized with a goat
anti-mouse IgG conjugated to FITC, and the anti-caveolin antibody was detected with a
goat anti-rabbit IgG linked to FITC. Staining from three independent experiments was
analyzed by deconvolution microscopy and shown to be reproducible. The data shown
represent analysis of 0.2 pm sections through the cells.

109

Figure 4-8. Intracellular staining of normal and LPI fibroblasts with antibodies against
proteins of the endoplasmic reticulum and Golgi Complex. According to the protocol
described in the Methods Chapter, normal (A, C, and E) and LPI (B, D, and F) fibroblasts
were fixed with -20C MeOH and subjected to immunohistochemistry with the anti-
KDEL antibody specific for the endoplasmic reticulum (A and B), the anti-
sialyltransferase (ST) antibody specific for the Golgi and Trans-Golgi Network (TGN) (C
and D), and the anti-mannose 6-phosphate receptor (M6PR) antibody specific for the
TGN and late endosomes (E and F). The anti-KDEL primary antibody was visualized
with a goat anti-mouse IgG conjugated to FITC, and the anti-ST and anti-M6PR
antibodies were detected with a goat anti-rabbit IgG linked to FITC. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. The data shown represent analysis of 0.2 pm sections through the cells.

Ill

Figure 4-9. Intracellular staining of normal and LPI fibroblasts with antibodies against
proteins of the endocytic and recycling pathways. According to the protocol described in
the Methods Chapter, normal (A and C) and LPI (B and D) fibroblasts were fixed with
-20C MeOH and subjected to immunohistochemistry with the anti-Rab5 antibody
specific for early endosomes (A and B), and the anti-transferrin receptor (TfR) antibody
specific for vesicles involved in recycling (C and D). The anti-TfR primary antibody was
visualized with a goat anti-mouse IgG conjugated to FITC, and the anti-Rab5 antibody
was detected with a goat anti-rabbit IgG linked to FITC. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
The data shown represent analysis of 0.2 pm sections through the cells.

113

Figure 4-10. Intracellular staining of normal and LPI fibroblasts with an antibody against
a lysosomal enzyme. According to the protocol described in the Methods Chapter,
normal (A) and LPI (B) fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with the anti-cathepsin D antibody specific for lysosomes and
related acidic vesicles. The anti-cathepsin D primary antibody was visualized with a goat
anti-rabbit IgG conjugated to FITC. Staining from three independent experiments was
analyzed by deconvolution microscopy and shown to be reproducible. Only one cell is
shown in panel A, whereas one of several cells is outlined in panel B in order to
distinguish the cell periphery. The data shown represent analysis of 0.2 pm sections
through the cells.

115

Figure 4-11. Intracellular staining of normal and LPI fibroblasts with an antibody against
a lysosomal membrane protein. According to the protocol described in the Methods
Chapter, normal (A) and LPI (B) fibroblasts were fixed with -20C MeOH and subjected
to immunohistochemistry with the anti-lpgl20 antibody specific for lysosomal
membranes. The anti-lpgl20 primary antibody was visualized with a goat anti-rabbit IgG
conjugated to FITC. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells.

117
15.O um

Figure 4-12. Visualization of acidic compartments of normal and LPI fibroblasts with
acridine orange. Normal (A) and LPI (B) fibroblasts were treated with acridine orange
(AO) according to the protocol described in Chapter 4. Briefly, cells were incubated with
5 pg/ml of AO for 15 min before fixing with -20C MeOH and viewing with a Nikon
Axiophot epifluorescence inverted microscope. The digitized images were captured
using a Spot CCD camera with a resolution of 1315 x 1033 pixels (Diagnostics
Instruments, Inc., Sterling Heights, MI). Fluorescence from three independent
experiments was analyzed and shown to be reproducible.

A

Figure 4-13. Lysosomal detection in normal and LPI fibroblasts following chloroquine
treatment. Normal (A and B) and LPI (C and D) fibroblasts were treated with
chloroquine and stained with anti-lpgl20 antibody according to the protocol described in
Chapter 4. Briefly, cells were incubated in MEM + 10% FBS (A and C) or MEM + 10%
FBS containing 50 mM chloroquine (B and D) for 1 h before fixing with -20C MeOH
and staining with anti-lpgl20 antibody. The anti-lpgl20 primary antibody was detected
with a goat anti-rabbit IgG linked to FITC. Staining from three independent experiments
was analyzed by deconvolution microscopy and shown to be reproducible. The data
shown represent analysis of 0.2 pm sections through the cells.

121

Figure 4-14. Lysosomal staining of normal and LPI cells expressing the GFP(C3)-CAT1
fusion protein. Normal (A) and LPI (B) fibroblasts were transfected for 3 h with the
GFP(C3)-CAT1 fusion protein according to the lipofectamine protocol described in the
Methods Chapter. Following 24-48 h of expression, cells were fixed with -20C MeOH
and stained with the anti-lpgl20 antibody. The lpgl20 primary antibody was detected
with a goat anti-rabbit IgG conjugated to Texas Red. Images were processed from three
independent transfections and the fluorescence patterns of expressed proteins were
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.

123

CHAPTER 5
CAVEOLAR COMPLEX BETWEEN THE CATIONIC AMINO ACID
TRANSPORTER 1 AND ENDOTHELIAL NITRIC OXIDE SYNTHASE
Introduction
The nitric oxide synthases (eNOS, iNOS, and nNOS) catalyze the conversion of
L-arginine to L-citrulline and nitric oxide (NO), a nitrogen-centered free radical with
multiple and unique physiologic and bioregulatory activities (reviewed by Moneada and
Higgs, 1993; Nathan and Xie, 1994). Pulmonary endothelial cells generate NO via the
catalytic action of an NADPH-requiring, Ca+7calmodulin-dependent NO synthase
(eNOS) that is membrane-associated (Moneada et al., 1991; Forstermann et ah, 1994;
Palmer et ah, 1988). Pulmonary endothelial cells are a rich source of nitric oxide (NO)
which functions as both a paracrine and autocrine mediator. As a paracrine mediator, NO
controls vascular smooth muscle tone and inhibits leukocyte adhesion and platelet
aggregation. As an autocrine mediator, NO influences growth factor signals and cellular
proliferation. Many cardiovascular diseases, including hypertension, diabetes,
atherosclerosis, and heart failure, develop complications because endothelial cells fail to
produce adequate amounts of NO.
In endothelial cells, eNOS-mediated formation of NO from arginine is dependent
upon an adequate and continuing supply of arginine (Aisaka et ah, 1989; Cooke et ah,
1991; Taylor and Poston, 1994; Rossitch et ah, 1991). Several studies have shown that
the half-saturating arginine concentration for eNOS is less than 10 pM (Liu et ah, 1996;
124

125
Palmer and Moneada, 1989; Pollock et al., 1991). It has been reported, by several
laboratories, that intracellular arginine concentrations range from 100 to 800 pM in
cultured endothelial cells (Block et al., 1995; Baydoun et al., 1990; Hecker et al., 1990;
Mitchell et ah, 1990; Gold et ah, 1989). Consequently, eNOS should be saturated in
these cells. Therefore, increasing the extracellular arginine should not increase NO
production any further. However, a number of in vitro and in vivo studies indicate that
NO production by vascular endothelial cells under physiological conditions can be
increased by extracellular arginine (Aisaka et ah, 1989; Cooke et ah, 1991; Taylor and
Poston, 1994; Rossitch et ah, 1991; Eddahibi et ah, 1992). Furthermore, a recent report
by Arnal et ah demonstrates that the intracellular concentration of arginine in endothelial
cells can be varied over 100-fold without changing NO production (Arnal et ah, 1995).
This observation, i.e., that extracellular arginine administration seems to drive NO
production even when intracellular levels of arginine are available in excess, has been
termed the arginine paradox and cannot be explained based on the available data (Kurz
and Harrison, 1997). One paradigm that would explain this observation is that in
endothelial cells the intracellular arginine is in one or more pools that are poorly, if at all,
accessible to eNOS, whereas extracellular arginine transported into the cell is
preferentially delivered to eNOS. Under this paradigm, a plasma membrane arginine
transporter might be in close spatial alignment with, or directly linked to, the plasma
membrane bound eNOS protein.
Arginine transport is mediated by several independent transport activities in
mammalian cells (White, 1985; Malandro and Kilberg, 1996; Closs, 1996). In porcine

126
pulmonary artery endothelial cells (PAEC), System y+ has been extensively characterized
(Zharikov and Block, 1997), and is responsible for 60-80% of total carrier-mediated
arginine uptake (Greene et al., 1993; McDonald et ah, 1997). In 1991, two laboratories
independently documented that the native biologic function of the previously cloned
murine ecotropic retroviral receptor was System y+ transport activity (Kim et ah, 1991;
Wang et ah, 1991). The mRNA and corresponding protein, termed CAT1, are expressed
in a wide variety of cells, with the notable exception of liver (Kim et ah, 1991; Wang et
ah, 1991; Wu et ah, 1994; Kakuda et ah, 1993). Using a CAT1 antibody generated in our
laboratory, it was shown that the arginine transporter is concentrated in specific regions
of the plasma membrane in a number of cell types, including PAEC (Woodard et ah,
1994; McDonald et ah, 1997).
Considerable information has been published regarding the presence of plasma
membrane micro-domains referred to as caveolae (Parton, 1996; Simionescu and
Simionescu, 1987; Schnitzer et ah, 1994; Anderson, 1993; Lisanti et ah, 1994). These
specialized membrane regions contain one or more of a family of structural proteins
called caveolins, as well as numerous signaling proteins, such as G-protein coupled
receptor systems, and a high cholesterol content. Of the three known isoforms, caveolin-
1 and caveolin-2 are most abundantly expressed in fibroblasts, endothelial cells, and
adipocytes, whereas caveolin-3 exhibits muscle-specific expression (reviewed by
Okamoto et ah, 1998). In addition to concentrating specific proteins within a specialized
region of the plasma membrane, caveolin may play a role in regulating the activation of
associated proteins. Caveolin homo- and heterodimers (between caveolin-1 and -2) form
a membrane-embedded hairpin structure resulting in cytosolic N- and C-termini (Scherer

127
et al., 1997). The caveolin scaffolding domain, which includes amino acids 82-101, has
been reported to interact with Ga subunits, Ha-Ras, Src family tyrosine kinases, and
eNOS (Song et al., 1996; reviewed by Okamoto et ah, 1998). With the exception of Ha-
Ras, a peptide encoding the scaffolding domain sequence can functionally inactivate each
of these proteins (Song et ah, 1996). The bradykinin and p-adrenergic receptors are
dependent on ligand binding for migration to caveolae, whereas the caveolar localization
of the endothelin receptor and amyloid precursor protein is ligand-independent (Chun et
ah, 1994; Feron et ah, 1997; de Weerd et ah, 1997).
eNOS has been shown to localize to the Golgi in cultured bovine aortic
endothelial cells (EC), human umbilical vein EC, and intact human blood vessels (Morin
and Stanboli, 1993; O'Brien et ah, 1995; Sessa et ah, 1995). In 1996, several laboratories
also documented significant localization of eNOS to plasmalemma caveolae of cultured
bovine lung microvascular EC and in luminal regions of plasmalemma isolated from
intact, perfused rat lungs (Garcia-Cardena et ah, 1996; Shaul et ah, 1996). For example,
Garcia-Cardena et ah showed that caveolin and membrane-bound eNOS co-localize in
lung microvascular endothelial cells and antibodies against one could be used to
immunoprecipitate the other, strongly suggesting that eNOS is complexed with caveolin
(Garcia-Cardena et ah, 1996). More recently, several labs have reported a predicted
caveolin-binding sequence (FSAAPFSGW) in the catalytic domain of eNOS that
interacts with caveolin-1 at residues 82-101. It is known that the aromatic residues are
essential for the recognition of caveolin-1, and when Sessa and colleagues mutated the
aromatic amino acids to alanine, caveolin no longer inhibited eNOS activity (Garcia-

128
Crdena, et al., 1997). Caveolin-bound eNOS is inactive until sufficient Ca2+-calmodulin
is present to compete for the caveolin binding site, thereby activating the enzyme to
produce NO. For previously unknown reasons, caveolar localization optimizes the ability
of eNOS to produce NO (Liu et al., 1996; Palmer and Moneada, 1989; Pollock et al.,
1991; Feron et al., 1996). The contribution of arginine transport to this phenomenon is
the subject of the research presented in this chapter.
An area under intense investigation concerns the mechanism by which specific,
yet unrelated, proteins are sequestered within the caveolae. In addition to sharing a
caveolin-binding motif, most of the proteins localized to caveolae are lipid modified.
There is significant evidence to suggest that acylation plays a role in targeting, although
the type and extent of the fatty acid requirement varies between different proteins. For
instance, eNOS is co-translationally myristoylated on glycine 2, and post-translationally
palmitoylated on cysteines 15 and 26. Fatty acylation moieties on these residues are
necessary, and sufficient, for targeting eNOS to caveolae (Garcia-Cardena et al., 1996).
The heterotrimeric G protein subunit az requires myristoylation at glycine 2 for stable
membrane association, and palmitoylation at cysteine 3 for specific localization to the
plasma membrane caveolae (Song et al., 1997; Morales et al., 1998). Myristoylation-
minus mutants of c-Src are poorly accumulated in caveolae, and other proteins, including
SNAP-25 and p59fyn, demonstrate a fatty acylation requirement for trafficking to the
plasma membrane, although caveolar localization has not been confirmed (Gonzalo and
Linder, 1998; van't Hof and Resh, 1997). However, in the case of p59fyn, it is believed
that protein synthesis and myristoylation occur on soluble ribosomes. This is followed by

129
rapid palmitoylation, plasma membrane association, and finally, partitioning to detergent-
insoluble membrane sub-domains, possibly caveolae.
The hypothesis of this project is that, in PAEC, the CAT1 transporter-containing
clusters, mentioned above, represent plasma membrane caveolae, and co-localization of
CAT 1-mediated arginine transport and eNOS provides an efficient mechanism for
delivery of substrate for NO synthesis, perhaps even in a direct manner. Additionally, the
trafficking of CAT1 to the caveolae may provide an alternative regulatory mechanism for
the production of NO. The following experiments were performed to test for the co
localization of CAT1 and eNOS within PAEC caveolae, as well as to determine if a
targeting signal for caveolar localization exists in the CAT1 protein sequence. My
working hypothesis is that the CAT1 transporter is co-localized with eNOS in caveolae
and thus, directs arginine to the enzyme for a more effective synthesis of NO.
Methods
Immunofluorescence assay. The immunofluorescence assays in this chapter were
performed in triplicate according to the methodology in Chapter 2. The antibodies used
in this chapter are described in Table 5-1. The secondary antibodies used were goat anti
rabbit or goat anti-mouse IgG conjugated to either FITC or TEXAS RED. The secondary
antibodies were used at a dilution of 1:200, and all immunofluorescence assays in this
chapter were analyzed by deconvolution microscopy.

130
Table 5-1
Antibodies for Immunofluorescence Studies
Name
Host
Source
Dilutions
IF* IB**
CAT1
rabbit
Dr. Michael Kilberg,
1:50 -
Univ. of Florida
caveolin
mouse
Transduction Laboratories
1:50 -
eNOS
mouse
Lexington, KY
Transduction Laboratories
1:10 1:200
Lexington, KY
(3-tubulin
mouse
Sigma,
1:100 -
St. Louis, MO
sialyltransferase
rabbit
Dr. William Dunn,
1:50 -
Univ. of Florida
10e6
mouse
Dr. William Brown,
1:100 -
Cornell Univ.
*IF = immunofluorescence
**IB = immunoblotting
Immunodepletion of CAT1 transport activity. By collaborators in the laboratory
of Dr. Edward Block, plasma membrane vesicles were prepared by sucrose gradient
centrifugation as described by Teitel (Teitel, 1986) and modified by Bhat and Block
(Bhat and Block, 1990; Bhat and Block, 1992). Plasma membrane proteins were
solubilized by the method described by Fafournoux et al. (Fafoumoux et al., 1989). The
solubilized proteins in the supernatant were precipitated by incubation with 20%
polyethylene glycol (PEG-8000) at 4C for 20 min. Immunodepletion of CAT1
transporter was performed using the protocol of Tamarappoo et al. (Tamarappoo et al.,
1992). Briefly, a 1-ml aliquot of goat anti-mouse IgG covalently linked to agarose beads
(Sigma Chemical Co., St. Louis, MO) was incubated for 1 h with 20 pg of a mouse anti
human eNOS antibody on ice and spun for 5 min at 1,500 x g, after which the supernatant
was discarded. The agarose beads were then washed once with STAB buffer (20%
glycerol, 2 mM EDTA, 2 mM DTT, 0.2% sodium cholate, 0.25% asolectin, and 10 mM

131
HEPES, pH 7.4) and mixed with solubilized proteins resuspended in STAB buffer. After
incubation for 1 h on ice, the beads were pelleted as above and discarded, the supernatant
was saved, and the solubilized, non-precipitated proteins were reconstituted into
proteoliposomes.
SDS-PAGE and anti-eNOS immunoblotting. A 125 pi aliquot of sample dilution
buffer (SDB) with (3-mercaptoethanol (BME) was added to the protein A-Sepharose
immunoprecipitated proteins. The samples were heated to 65C for 10 min and then the
sepharose was pelleted by spinning at 13,000 x g for 10 min. Supernatant was removed
from the top and placed in a centrifuge tube and heated again for 10 min at 65C. Each
sample (supernatant from above) was loaded in one lane and subjected to one
dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE)
(Laemmli, 1970; Chiles et al., 1987) and transferred at 299 mAmps for 18 h to a
nitrocellulose membrane (Chiles et al., 1987). The nitrocellulose membrane was blocked
with 5% non-fat dry milk (NFDM) at room temperature for 1.5 h, rinsed with
TBS/TWEEN (30 mM Tris base, 150 mM NaCl, 0.1% Tween 20, pH 7.6), then incubated
with primary antibody prepared in TBS/TWEEN for 1-2 h at room temperature. Two
quick rinses, one 15 min rinse, and two 5 min rinses with TBS/TWEEN were followed by
incubation of the nitrocellulose in the secondary antibody, prepared in TBS/TWEEN, for
1-2 h at room temperature. After washing nitrocellulose six times for 5 min each, the blot
was incubated in 6 ml of a 1:1 mixture of the Enhanced Chemiluminescence (ECL)
reagents (Pierce, Rockford, IL) for 1 min, drained, wrapped in plastic, and immediately
exposed to film.

132
Reconstitution and assay of amino acid transport. Reconstitution of soluble
proteins into proteoliposomes was performed following the protocol of Fafoumoux et al.
(Fafoumoux et al., 1989) and transport assays were performed as described previously
(Zharikov and Block, 1997). Briefly, plasma membrane vesicles or proteoliposomes (20
pg/30 pi) in SMB buffer (250 mM sucrose, 1 mM MgCl2, 10 mM Flepes, pH 7.5 ) were
added to 270 pi of external solution containing 140 mM NaSCN, 1 mM MgS04, 10 mM
HEPES-Tris (pH 7.4), and 50 pM [3H]-arginine or 50 pM [3H]-glutamine. After
incubation for 3 min at 37C, reactions were terminated by the addition of 5 ml of ice-
cold 140 mM NaCl (stop solution) followed by filtration through glass-fiber Whatman
GF/C filters presoaked in 0.3% polyethylenimine to decrease the nonspecific absorption
of [3H]-L-arginine or glutamine. The filters were washed 4 times with 5 ml of stop
solution, dried, and trapped radioactivity determined using liquid scintillation
spectrometry. Zero-time blank values (membrane vesicles or proteoliposomes added
after stop solution) were subtracted from all experimental values. The data represent
triplicate assays performed on at least two independent preparations, and are presented as
pmol.mgTprotein.3min'1. Statistical significance was established by Students T test.
Results
Characterization of CAT1 staining of PAEC plasma membranes. Using a CAT1
antibody, generated in our laboratory, in conjunction with immunostaining analysis using
an epifluorescence microscope, it was previously shown that the arginine transporter is
concentrated in specific regions of the plasma membrane in a number of cell types,
including PAEC (Woodard et al., 1994). These results were confirmed and extended

133
using deconvolution microscopy as a means of obtaining high-resolution data analysis
(McDonald et al., 1997). The CAT1 transporter is not uniformly distributed over the
entire cell surface. Instead, the cell surface has discrete regions that contain a high
transporter content. Figure 5-1 (A and B) shows PAEC stained with a 1:50 dilution of
anti-CATl antibody detected with a 1:200 dilution of goat anti-rabbit IgG linked FITC.
Staining with the anti-CATl transporter antibody was completely inhibited by
preadsorption of the antibody with the corresponding peptide antigen (Fig. 5-1 C),
whereas incubation of the antibody with non-antigen peptide sequences from within the
CAT1 transporter did not block cell staining (data not shown).
Co-localization of CAT1 and caveolin within PAEC. To determine if the
observed clustering is a result of CAT1 localization to caveolae, PAEC were fixed with a
4% paraformaldehyde solution and stained with a 1:50 dilution of anti-CATl (Figure 5-2
A), a 1:50 dilution of anti-caveolin (Figure 5-2 B), or a combination of both antibodies
(Figure 5-2 C). The CAT1 antibody was detected using a 1:200 dilution of goat anti
rabbit IgG conjugated to FITC, and the caveolin antibody was visualized with a 1:200
dilution of goat anti-mouse IgG linked to Texas Red. Immunohistochemistry of the
PAEC with antibodies specific for caveolin documented intensely staining clusters,
consistent with proposed caveolae structure and localization. Co-staining of the PAEC
with anti-CATl and anti-caveolin antibodies resulted in significant overlap with regard to
the localization of the two proteins (Figure 5-2 C). It is clear from these results that the
majority of the clusters containing the arginine transporter coincide with caveolae in the
PAEC.

134
Co-localization of CAT1 and eNOS within PAEC. Several laboratories have
documented the co-localization of caveolin and eNOS by immunohistochemistry, and
shown that eNOS is co-immunoprecipitated with anti-caveolin antibodies (Garcia-
Cardena, 1996; Feron et al., 1996). When these co-localization experiments were
reproduced by myself using PAEC and human fibroblasts, caveolin and eNOS showed
significant overlap in both cell types (data not shown). These results, in addition to the
presence of CAT1 in the caveolae, raised the possibility that CAT1 and eNOS may be co
localized in a caveolar complex to facilitate NO synthesis. When paraformaldehyde-
fixed PAEC were subjected to immunohistochemistry to test for co-localization of CAT1
and eNOS, a significant portion of the detectable membrane-associated eNOS was
clustered within specific regions on the cell surface (Figure 5-3 C). The CAT1 antibody
was detected with a goat anti-rabbit IgG conjugated to FITC, and eNOS was visualized
with a goat anti-mouse IgG linked to Texas Red. Although clearly separate pools also
exist, many of the eNOS-labeled membrane clusters overlapped with the CAT1 staining.
A large quantity of eNOS and CAT1 were detected in a perinuclear region
resembling the Golgi. To confirm that the perinuclear staining was Golgi, both eNOS
and CAT1 were used in double-labeling experiments with Golgi-specific antibodies.
When MeOH-fixed PAEC were labeled with a 1:50 dilution of CAT1 and a 1:100
dilution of the cis-Golgi marker, 10e6 (see Table 5-1), the CAT1 antibody demonstrated
strong co-localization with 10e6 in the Golgi region (Figure 5-4 A). Likewise, eNOS
(1:10 dilution) and the Golgi-marker, sialyltransferase (1:50 dilution), showed significant
overlap in the perinuclear region of MeOH-fixed PAEC (Figure 5-4 B). These data are
consistent with the report by Sessa et al. that a significant amount of eNOS resides in the

135
Golgi (Sessa et al., 1995). Although there appears to be some co-localization of CAT1
and eNOS in the Golgi region, it is uncertain as to whether an interaction is already
present within the Golgi, or if the two proteins are just in close proximity as they advance
along the biosynthetic or recycling pathways. Further research is needed to clarify the
site of formation and the potential intracellular localization of the complex.
Paraformaldehyde-fixed PAEC were also dual stained with eNOS and a 1:50 dilution of
GLUT1 (see Table 4-1) or a 1:100 dilution of (3-integrin (see Table 4-1) to determine if
membrane proteins other than CAT1 co-localized with eNOS. Results from three
separate experiments demonstrated no significant overlapping fluorescence in either case
(data not shown).
Transfection of CAT1-GFP in PAEC. To confirm the results obtained from
immunofluorescence studies to detect endogenous CAT1, as well as to develop an
expression system to be used in future experiments, a GFP(C3)-CAT1 fusion protein was
constructed with the GFP tag at the N-terminus of the CAT1 sequence (see Methods
Chapter for details). PAEC were transfected with GFP(C3)-CAT1 or GFP(C3) vector
alone according to the Lipofectamine protocol (see Methods Chapter). Following 24 to
48 h of expression, PAEC were fixed with -20C MeOH and visualized by deconvolution
microscopy. In the absence of a targeting signal, the GFP(C3) shows diffuse
fluorescence throughout the cytoplasm and nucleus (data not shown). The GFP(C3)-
CAT1, on the other hand, is targeted specifically to the plasma membrane, and apparently
vesicle compartments involved in biosynthesis and intracellular trafficking as indicated
by a strong punctate pattern throughout the cytoplasm (Figure 5-5 A). Although the over-

136
expression resulted in large quantities of GFP(C3)-CAT1 on the plasma membrane, there
were still specific regions of more highly concentrated transporter. The CAT1 antibody
(1:50 dilution) showed strong co-localization with the expressed GFP(C3)-CAT1 in
PAEC, as illustrated in Figure 5-5 B. However, there were separate pools of unstained
GFP(C3)-CAT1 fusion protein and CAT1 antibody (detected by a 1:200 dilution of goat
anti-rabbit IgG linked to Texas Red). This was probably due to CAT1 antibody staining
of endogenous transporter, as well as a less than 100% efficiency of CAT1 antibody
staining of the expressed GFP(C3)-CAT1 protein. To determine if the GFP(C3)-CAT1
co-localized with eNOS in the caveolae, PAEC were transfected with the fusion protein,
fixed with -20C MeOH, and stained with caveolin (Figure 5-6 A) or eNOS (Figure 5-6
B) antibodies. Both antibodies were visualized using goat anti-mouse IgG conjugated to
Texas Red. Although significant co-localization of GFP(C3)-CAT1 with caveolin or
eNOS existed, the over-expression also resulted in a greater quantity of transporter that
did not appear to exist in caveolae.
Co-localization of CAT1 and eNOS following nocodazole treatment. As
mentioned previously, the CAT1 clusters can be dispersed by treatment with the
microtubule-destabilizing drug, nocodazole (Woodard et al., 1994). Several reports have
suggested that caveolin also relies partially on microtubules for cycling between the
Golgi and the plasma membrane caveolae (Conrad et al., 1995). To determine whether or
not CAT1 and eNOS are dependent on microtubules for co-localization, PAEC were
treated with MEM + 25 pg/ml nocodazole (prepared from a 3mg/ml nocodazole stock in
DMSO) for 1 h at 37C before double-staining with anti-CATl and anti-eNOS antibodies

137
(Figure 5-7 D), as described above. Cells were incubated with MEM + DMSO (0.83%),
as illustrated in Figure 5-7 C, to ensure that the DMSO was not responsible for effects
observed with nocodazole treatment. PAEC were also stained with a 1:100 dilution of
mouse anti-rat P-tubulin antibody following either DMSO (Figure 5-7 A) or nocodazole
(Figure 5-7 B) treatment to test for microtubule disruption. The tubulin was completely
disrupted following the 1 hr nocodazole treatment. This was determined based on the
shift of the fluorescent pattern from long, thin strands of microtubules to small punctate
free tubulin subunits. Five to ten images, collected from three individual experiments
indicated that the eNOS Golgi staining and the CAT1 "patches" were altered in the
nocodazole-treated cells. eNOS staining in the perinuclear region was disrupted, and the
eNOS-labeled vesicles were more diffusely distributed throughout the cytoplasm. The
CAT1 staining disappeared completely from the perinuclear region, and the "patches"
appeared to be smaller and almost entirely located at the cell periphery following
nocodazole treatment. The co-localization of CAT1 and eNOS in DMSO appeared to be
predominantly intracellular, and was completely disrupted following nocodazole
treatment. A sample of the PAEC were allowed to recover in MEM + 10% FBS for 3 h
after removing the nocodazole. These cells were then fixed with 4% paraformaldehyde
and stained with antibodies against CAT1 and eNOS. Three hours is long enough for the
re-polymerization of microtubules, and this was confirmed by staining nocodazole-treated
PAEC with anti-(3-tubulin antibody after the 3 h recovery period (data not shown). The
reformation of the CAT1 clusters, as well as co-localization of CAT1 and eNOS,
appeared to be complete after a 3 h recovery in MEM + 10% FBS (data not shown).

138
Immunodepletion of CAT1 activity with an eNOS-specific antibody. Co
localization of the CAT1 arginine transporter and eNOS within caveolae is consistent
with the proposal that these membrane micro-domains are a site for concentrating
proteins involved in signaling (Parton, 1996; Simionescu and Simionescu, 1987;
Schnitzer et al., 1994; Lisanti et al., 1994). However, the present observations also raise
the intriguing possibility that the CAT1 arginine transporter and eNOS are physically
associated. Such a complex might provide a mechanism for directed delivery or even
channeling of newly acquired extracellular arginine to eNOS for NO synthesis. This
channeling would explain the "arginine paradox" whereby eNOS apparently
preferentially accepts extracellular arginine in the presence of saturating intracellular
substrate levels (Kurz et al., 1997). The saturable Na+-independent arginine transport
activity in PAEC vesicles has been characterized previously and is mediated primarily by
the CAT1 transporter given that neither b0,+ nor y+L are detectable in these assays
(Zharikov and Block, 1997). It has also been determined that PAEC contain CAT1, but
little or no CAT2 or CAT2a mRNA (unpublished results). As a test for a direct complex
between CAT1 and eNOS, PAEC plasma membrane arginine transport activity was
detergent solubilized and subjected to immunoprecipitation with anti-eNOS antibody, as
described in the Methods Chapter (Fafournoux et al., 1989). Using the proteins in the
immunodepleted supernatant to reconstitute proteoliposomes (Tamarappoo et al., 1992)
provided an assay to check for anti-eNOS dependent immunodepletion of arginine
transport. Immunodepletion with non-immune mouse IgG caused no loss of
reconstitutable arginine transport, whereas the anti-eNOS monoclonal antibody caused
immunoprecipitation of 73% of the NaMndependent arginine transport (Table 5-2).

139
Table 5-2
Immunodepletion of CATI-mediated Arginine Transport
Activity by anti-eNOS Antibody
Immunoprecipitation
Prior to Reconstitution
Transport Velocity
pmol mg-1 protein 3
Percent of Control
min'
Arsinine Transport
No antibody
3511 50

Mouse IgG
4084 + 44
100
Anti-eNOS IgG
1113 17*
27
Glutamine Transport
Mouse IgG
4026176
100
Anti-eNOS IgG
4368 706
109
* P < 0.001 versus no antibody
or non-immune mouse IgG, i
n = 3 independent
experiments
Table 5-2. After solubilization of CAT1 transport activity (Fafournoux et al., 1989), the
protein fraction was subjected to immunodepletion (Tamarappoo et al., 1992), and then
incubated with either control mouse IgG or anti-eNOS IgG bound to anti-mouse IgG
immobilized on agarose beads. After pelleting the beads, proteins remaining in the
supernatant were reconstituted into proteoliposomes to test for immunodepletion of
saturable, Na+-independent arginine transport activity (Zharikov and Block, 1997). As a
control for transporter specificity by immunodepletion, the reconstituted proteoliposomes
also were assayed for Na+-dependent 50 pM glutamine transport to demonstrate that not
all amino acid transporters were precipitated by the eNOS antibody.
To establish that the anti-eNOS did not result in immunodepletion of arginine transport in
a non-specific manner, glutamine transport was monitored in the reconstituted
proteoliposomes after immunoprecipitation with non-immune and anti-eNOS IgG. No
immunodepletion of glutamine transport was observed. These data suggest that a
functional association exists between the CAT1 arginine transporter and membrane-
bound eNOS in PAEC.
Immunoprecipitation of eNOS using anti-CATl antibody. As mentioned
previously, several laboratories have documented the co-immunoprecipitation of eNOS
and caveolin in endothelial cells (Garcia-Cardena et al., 1996). To provide additional

140
evidence for a direct interaction between CAT1 and eNOS, CAT1 antibody-coated
agarose beads were generated as described above for binding of anti-eNOS to the beads
and used to immunoprecipitate CAT1 from detergent solubilized PAEC plasma
membrane vesicles. As controls, non-coated or non-immune IgG-coated agarose beads
were used in the immunoprecipitation. Proteins recovered from pelleted beads, as well as
the supernatant, were run on a SDS-PAGE gel and immunoblotted with an anti-eNOS
antibody (Figure 5-8). eNOS was detected in both the supernatants from control
immunoprecipitations as well as immunoprecipitations with the CAT1 antibody, although
less protein appeared in the latter supernatant. When the immunoprecipitates were
immunoblotted with anti-eNOS antibody, eNOS protein was detected primarily in the
control immunoprecipitation presumably the result of non-specific adsorption, with little
protein in the sample immunoprecipitated with the CAT1 antibody. This experiment was
repeated five times in order to optimize the conditions and refine the technique, however,
each attempt at co-precipitating eNOS with the anti-CATl antibody was unsuccessful.
Covalently cross-linking by treating PAEC membrane vesicles with dimethyl
suberimidate (DMS) prior to immunoprecipitation with the anti-CATl antibody was
performed. Chemical cross-linking reagents introduce covalent bonds between
neighboring proteins that are tightly associated. DMS, specifically, is one of a group of
chemicals that reacts with the e-amino groups of lysines and available N-terminal amines
(Bu and Schwartz, 1994). The cross-linking was attempted in the event that the CAT1-
eNOS interaction was being disrupted during the immunoprecipitation procedure. The
same immunoprecipitation procedure was followed (see Methods Section of this chapter),

141
with the addition of a 1 h incubation of PAEC membrane vesicles with 5 mM DMS prior
to the solubilization of plasma membrane proteins. Similar results, in which no co
precipitation of eNOS with CAT1, were obtained from immunoprecipitations performed
after treatment with the cross-linking reagent, DMS (data not shown).
Mutagenesis of potential caveolar targeting sequences in CAT1. The following
experiments were performed to investigate the mechanism by which CAT1 is localized to
caveolae. The Introduction to this chapter discusses a variety of proteins that are subject
to myristoylation, palmitoylation, or a combination of both as targeting signals for
caveolar localization. CAT1 has a glycine at position 2 and three cysteines at positions 3,
20, and 30 that are conserved between the mouse and human transporter, and, as
described previously, corresponding residues in other proteins, including eNOS, are
required for caveolar localization. To determine if these residues are necessary for
targeting CAT1 to caveolae, three mutants were constructed (see Methods Chapter), the
cDNAs expressed in PAEC, and cells assayed for co-localization with anti-caveolin and
anti-eNOS antibodies. The following mutations were made using the GFP(C3)-CAT1 as
the template (see Methods Chapter), and confirmed by restriction analysis and sequencing
(ICBR DNA Sequencing Core): CATMUT1, cysteines at 20 and 30 changed to serines;
CATMUT2, glycine at position 2 was changed to alanine; CATMUT3, glycine at
position 2 changed to alanine and cysteines at positions 3, 20, and 30 changed to serine.
Wild-type GFP(C3)-CAT1 and each of the mutants were transfected into PAEC as
described in the Methods Chapter, fixed with -20C MeOH after 24 to 48 h of expression,
and stained with anti-caveolin or anti-eNOS antibodies. The caveolin and eNOS
antibodies were detected using a goat anti-mouse IgG conjugated to Texas Red. As

142
demonstrated earlier, both the caveolin and eNOS antibodies co-localize with the
expressed wild-type GFP(C3)-CAT1 (Figure 5-6). However, the GFP(C3)-CAT1 is
expressed at a much higher level than the endogenous, therefore, there is a significant
amount of the fusion protein on the plasma membrane that does not co-localize with
either caveolin or eNOS. CATMUT1 was expressed normally, trafficked to the plasma
membrane, and co-localized with both caveolin (Figure 5-9 A) and eNOS (Figure 5-9 B)
to the same degree as the wild-type fusion protein. CATMUT2 and CATMUT3 also
expressed well and were detected on the plasma membrane, however, CATMUT3
showed very little co-localization with either caveolin (Figure 5-10 A) or eNOS (Figure
5-10 B). It is hard to determine whether or not there is a true difference between the
wild-type and CATMUT3, though, because the amount of co-localization following the
transfections was low in the wild-type. Also, the transfected cells had a reduced amount
of eNOS on the plasma membrane, so the majority of the co-localization between
GFP(C3)-CAT1 and eNOS was in the perinuclear region. Antibodies against caveolin
and eNOS inconsistently co-localized with the expressed CATMUT2 protein, in that,
some cells would detect overlapping of the two signals and others would not (data not
shown). Whereas this may indicate an intermediate result, no conclusive statements
following three individual experiments can be made concerning the CATMUT2
construct.
eNOS distribution following exposure to extracellular L-arginine. As illustrated
in Figure 5-11, a significant amount of eNOS resides in the Golgi of PAEC. However,
for enzyme activation and maximal NO production, eNOS must be translocated to the
caveolar regions (Liu et al., 1996; Feron et al., 1996). Acylation is likely involved in

143
specifically directing eNOS to the caveolae, however, other mechanisms may be involved
in stimulating the trafficking when there is a need for NO production. Unpublished
observations in the laboratory of Dr. Edward Block, University of Florida, suggest that
reducing the amount of extracellular L-arginine reduces the amount to NO produced by
the PAEC. Also, preliminary results have shown that blocking the uptake of L-arginine,
using lysine as a competitive inhibitor, blocks the production of NO. To investigate
whether or not extracellular arginine concentrations detectably influence the amount of
eNOS on the plasma membrane, PAEC were incubated in complete MEM + 10% FBS
(Figure 5-11 A), complete MEM minus L-arginine + 10% FBS (Figure 5-11 B), or
complete MEM plus 0.5 mM L-arginine + 10% FBS (Figure 5-11 C). One set of cells
was incubated in the different media solutions for 1 h (under normal culture conditions,
see Methods Chapter), and a second set of cells was incubated overnight. Cells were then
fixed with -20C MeOH and stained with anti-eNOS antibody. Regardless of the
concentration of L-arginine in the media, the eNOS antibody staining was identical, with
a small amount of eNOS detected on the plasma membrane and abundant perinuclear
staining. These results suggest that extracellular L-arginine is not a factor in the
recruitment of eNOS from the Golgi to the plasma membrane in cultured PAEC.
Discussion
The clinical importance of the vascular L-arginine/nitric oxide synthase cycle is
validated when considering the number of pathophysiological conditions that arise from a
disruption of the system. Alterations in the production or availability of NO have been
implicated in contributing to complications arising in such diverse disorders as

144
atherosclerosis, diabetes, hypercholesterolemia, hypertension, aging, cigarette smoking,
and heart failure (Harrison, 1996; reviewed by Harrison, 1997). Little is known regarding
the mechanisms of NO action underlying the above medical conditions. An alteration in
the expression of eNOS is one factor that may exacerbate a disease state. Although
eNOS is constitutively expressed, it is subject to regulation by vascular shear stress
(Nishida et al., 1992), exposure to lysophosphatidylcholine (Zembowicz et ah, 1995), low
concentrations of oxidized low density lipoprotein (Hirata et ah, 1995), and cyclic GMP
analogues (Ravichandran and Johns, 1995). During its life cycle, eNOS is dually-
acylated, phosphorylated at multiple sites, and interacts with caveolin and
calcium/calmodulin, if not several other proteins. A breakdown in one or more of these
processes would affect the production and release of NO. Other factors that may lead to
abnormal NO production include a change in the availability of cofactors, such as
tetrahydrobiopterin, or a destruction of NO by reactive oxygen species.
It is believed that the availability of the substrate, L-arginine, contributes to the
regulation of NO production, however conflicting data regarding the effectiveness of L-
arginine treatment exist. Mugge and Harrison reported, in isolated organ chambers, that
L-arginine had no effect on the vasodilation of aortas from normal and cholesterol-fed
rabbits (Mugge and Harrison, 1991). However, numerous studies have shown positive
vascular responses to L-arginine treatment in experimental animals and humans with
hypercholesterolemia, diabetes, and hypertension (Cooke et al., 1991; Creager et al.,
1992). Oral administration of L-arginine doubles the plasma levels of the amino acid and
has been effective in treating hypertensive rats, as well as, atherosclerosis in cholesterol-
fed rabbits (Cooke et al., 1992; Chen and Sanders, 1993). Although data concerning the

145
responses to L-arginine have been collected for a number of diseases, in a variety of
experimental models, the mechanisms underlying the results remain elusive. L-arginine
may work directly on the endothelium through NO as a vasodilator, or it may block the
actions of endogenous NOS antagonists such as asymmetric dimethyl arginine (ADMA)
(Boger et al., 1997). This chapter presents evidence to support the concept that a
functional relationship, between eNOS and the CAT1 arginine transporter, provides
several additional potential regulatory mechanisms for the production of NO.
Immunofluorescence studies to detect endogenous proteins, as well as transfection
of a GFP(C3)-CAT1 construct, showed that caveolin, CAT1, and eNOS co-localized to
specific regions within the PAEC. In addition, an eNOS-specific antibody was able to
immunoprecipitate CAT 1-mediated transport activity. These data document a close
spatial alignment, and even suggest but do not prove a direct binding between these
proteins. Several laboratories have shown a direct interaction between caveolin-1 and
eNOS by reciprocal immunoprecipitations and yeast two-hybrid systems. However, our
attempts to immunoprecipitate eNOS with an antibody against CAT1 were unsuccessful,
even after cross-linking reagents were used to prevent dissociation during somewhat
harsh isolation conditions. There are numerous possible reasons for these negative
results. It is possible that the CAT1 antibody, although able to detect the transporter
during immunostaining protocols, was unable to recognize or interact with CAT1 bound
to eNOS following solubilization of the complex. The results from the PAEC
reconstituted vesicle transport following immunodepletion provide evidence for a direct
interaction between CAT1 and eNOS. The attempt to detect eNOS on an immunoblot
after immunoprecipitating with the CAT1 antibody would have provided additional

146
evidence, but those experiments were negative. It is possible that the eNOS antibody is
better suited for the immunoprecipitation protocol than the CAT1 antibody. The anti-
CAT 1 does not immunoblot effectively either. In the future, problems involving the
CAT1 antibody may be overcome by overexpressing the GFP(C3)-CAT1 in PAEC, and
then immunoprecipitating eNOS using an anti-GFP antibody. The CAT1 transporter may
be in such low abundance that the amount precipitated with eNOS is below the level of
detection of the immunoblot (especially, if the interaction between CAT1 and eNOS is
transient). Michel et al. proposed a model of reciprocal regulation in which eNOS binds
caveolin and the calcium/calmodulin complex via the same or overlapping sequences
(Michel et al., 1997). eNOS is inactive when it is bound to caveolin. Intracellular
concentrations of calcium are increased following agonist stimulation, allowing calcium
to bind, and activate, calmodulin. The calcium/calmodulin replaces caveolin and eNOS
is activated. It is possible that activated eNOS is allowed to diffuse freely within the
caveolar domain, interact transiently with CAT1 in order to produce NO, then return to
caveolin and the inactive state. Alternative approaches may address these questions. The
yeast two-hybrid system has proven to be a sensitive assay for detecting interactions
between proteins that fail to co-immunoprecipitate and could be applied to CAT1 and
eNOS. For instance, Ras and Raf did not co-precipitate together; however, an association
was detected using the yeast two-hybrid system (Van Aelst et al., 1993). The Kilberg
laboratory has initiated a collaboration with Dr. Bettie Sue Masters, Southwestern
Medical School, to test for binding between CAT1 peptide sequences and purified eNOS.
Association of the CAT1 arginine transporter and eNOS in PAEC provides a
mechanism for the directed delivery of L-arginine substrate to eNOS. NO is a highly

147
labile free radical that must reach the target tissue within approximately 10 seconds. The
CATl-eNOS caveolar machinery would provide the necessary degree of efficiency and
regulation for proper NO signaling. If this model holds, it would represent the first
example, in mammalian cells, of a functional complex between an amino acid transport
protein and an enzyme. Although a direct interaction has not been proven,
immunofluorescence and vesicle transport data indicate that a close spatial alignment and
functional relationship exist between CAT1 and eNOS in the caveolae. Directed delivery
of extracellular arginine to eNOS would account for the arginine paradox described
earlier (Kurz and Harrison, 1997) and would also explain the observation by Liu et al.
that caveolar localization of eNOS is required for optimal NO production by eNOS (Liu
et al., 1996). Understanding the subcellular environment and events that are required for
NO production in endothelial cells will provide a better understanding of the mechanisms
that regulate blood pressure in the pulmonary circulation of patients with lung disease.

Figure 5-1. Surface labeling of PAEC with the CAT1 transporter antibody. PAEC were
fixed with 4% PFA and subjected to immunohistochemistry using the procedures
described in the Methods Chapter. Panels A and B demonstrate the clustering of the
CAT1 transporter on PAEC using a polyclonal anti-CAT 1 antibody detected with a goat
anti-rabbit IgG conjugated to FITC. Panel A shows staining of PAEC before
deconvolution to reveal the outline of the entire cell body. Panel B shows the same cells
after deconvolution. Panel C demonstrates cells that were incubated with anti-CAT 1
antibody following preadsorption for 12 h at 4C with 50 pg/ml of the corresponding
peptide antigen. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells. One cell in panel C is outlined to
distinguish the cell periphery.

149

Figure 5-2. Co-localization of CAT1 and caveolin on PAEC. According to the
immunofluorescence protocol described in the Methods Chapter, PAEC were fixed with
4% PFA and stained with antibodies against CAT1 (A), caveolin (B), or both (C). Co
localization was assayed by simultaneously incubating PAEC with a polyclonal antibody
against CAT1 detected with goat anti-rabbit IgG linked to FITC and a monoclonal
antibody against caveolin detected with goat anti-mouse IgG conjugated to Texas Red.
Staining from three independent experiments was analyzed by deconvolution microscopy
and shown to be reproducible. The data shown represent analysis of 1.0 pm sections
through the cells.

151

Figure 5-3. Co-localization of CAT1 and eNOS on PAEC. According to the
immunofluorescence protocol described in the Methods Chapter, PAEC were fixed with
4% PFA and stained with antibodies against CAT1 (A), eNOS (B), or both (C). Co
localization was assayed by simultaneously incubating PAEC with a polyclonal antibody
against CAT1 detected with goat anti-rabbit IgG linked to FITC and a monoclonal
antibody against eNOS detected with goat anti-mouse IgG conjugated to Texas Red.
Staining from three independent experiments was analyzed by deconvolution microscopy
and shown to be reproducible. The data shown represent analysis of 1.0 pm sections
through the cells.

153
4.0 urn
4.0 uin

Figure 5-4. Detection of CAT1 and eNOS in the Golgi of PAEC. According to the
immunofluorescence protocol described in the Methods Chapter, PAEC were fixed with
-20C MeOH and co-stained with antibodies against CAT1 and 10e6 (A) or eNOS and
sialyltransferase (ST) (B). The anti-10e6 and anti-ST are antibodies specific for proteins
of the Golgi Complex. The anti-CAT 1 and anti-ST were detected with goat anti-rabbit
IgG linked to FITC, whereas anti-eNOS and anti-10e6 were labeled with goat anti-mouse
IgG linked to Texas Red. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells.

155

Figure 5-5. Expression of the GFP(C3)-CAT1 fusion protein in PAEC. PAEC were
transfected for 3 h with the GFP(C3)-CAT1 fusion protein according to the lipofectamine
protocol described in the Methods Chapter. Following 24-48 h of expression, cells were
fixed with -20C MeOH and visualized by deconvolution microscopy. Panel A shows
the expression pattern of the GFP(C3)-CAT1 fusion protein on the plasma membrane and
in intracellular vesicles. PAEC in panel B were stained with the CAT1 antibody
following expression of the GFP(C3)-CAT1 fusion protein. The CAT1 antibody was
visualized with a goat anti-rabbit IgG linked to Texas Red. Images were processed from
three independent transfections and the fluorescence patterns of expressed proteins were
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.

157

Figure 5-6. Immunofluorescent staining of PAEC transfected with GFP(C3)-CAT1
fusion protein. PAEC were transfected for 3 h with the GFP(C3)-CAT1 fusion protein
according to the lipofectamine protocol described in the Methods Chapter. Following 24-
48 h of expression, cells were fixed with -20C MeOH and stained with antibodies
against caveolin (A) and eNOS (B). Both the caveolin and eNOS antibodies were
detected with a goat anti-rabbit IgG conjugated to Texas Red and visualized by
deconvolution microscopy. Images were processed from three independent experiments
and the staining was determined to be reproducible. The data shown represent analysis of
1.0 pm sections through the cells. One cell in panel B is outlined in order to distinguish
the PAEC periphery.

159

Figure 5-7. Distruption of CATl/eNOS co-localization in PAEC treated with
nocodazole. PAEC were incubated with DMSO (A and C) or nocodazole (B and D) for 1
h then fixed with -20C MeOH and stained with an antibody against (3-tubulin (A and B)
or co-stained with antibodies against CAT1 and eNOS (C and D). The (3-tubulin and
eNOS antibodies were detected with goat anti-mouse IgG conjugated to FITC and the
CAT1 antibody was detected with goat anti-rabbit IgG linked to Texas Red. Images were
processed from three independent experiments and the staining was determined to be
reproducible. The data shown represent analysis of 1.0 pm sections through the cells.

161

162
A B
* 140 kDa
Figure 5-8. CAT1 immunoprecipitation of eNOS from solubilized PAEC plasma
membrane vesicles. The CAT1 antibody was used to immunoprecipitate eNOS
according to the protocol in the Methods Section of Chapter 5. Lane A shows eNOS
pelleted by control IgG linked to protein A-sepharose beads, presumably through non
specific binding, and lane B shows eNOS immunoprecipitated with the protein A-
Sepharose beads conjugated to the CAT1 antibody. Immunoblot analysis was performed
with a 1:200 dilution of eNOS antibody detected by a 1:20,000 dilution of goat anti
mouse IgG conjugated to horseradish peroxidase. Data shown are representative of five
individual blots.

Figure 5-9. Immunofluorescent staining of PAEC transfected with the CATMUT1
palmitoylation mutant. PAEC were transfected for 3 h with the CATMUT1 fusion
protein that has serine substitutions for cysteine at residues 20 and 30 and was prepared
according to the mutagenesis protocols described in the Methods Chapter. Transfection
was performed as outlined in the Methods Chapter. Following 24-48 h of expression,
cells were fixed with -20C MeOH and stained with antibodies against caveolin (A) and
eNOS (B). Both the caveolin and eNOS antibodies were detected with a goat anti-rabbit
IgG conjugated to Texas Red and visualized by deconvolution microscopy. Images were
processed from three independent experiments and the staining was determined to be
reproducible. The data shown represent analysis of 1.0 pm sections through the cells.

164

Figure 5-10. Immunofluorescent staining of PAEC transfected with the CATMUT3
palmitoylation mutant. PAEC were transfected for 3 h with the CATMUT3 fusion
protein that has serine substitutions for cysteine at residues 3, 20, and 30, and an alanine
substitution for glycine at residue 2, and was generated according to the mutagenesis and
transfection protocols described in the Methods Chapter. Transfection of the PAEC was
performed according to the procedures described in the Methods Chapter. Following 24-
48 h of expression, cells were fixed with -20C MeOH and stained with antibodies
against caveolin (A) and eNOS (B). Both the caveolin and eNOS antibodies were
detected with a goat anti-rabbit IgG conjugated to Texas Red and visualized by
deconvolution microscopy. Images were processed from three independent experiments
and the staining was determined to be reproducible. The data shown represent analysis of
1.0 pm sections through the cells.

166

Figure 5-11. eNOS staining of PAEC treated with varying concentrations of extracellular
L-arginine. PAEC were incubated with MEM + 10% FBS (A), MEM minus L-arginine
+ 10% FBS (B), and MEM + 0.5 mM L-arginine + 10% FBS (C) for 12 h under normal
culture conditions (see Methods Chapter). Cells were then fixed with -20C MeOH and
stained with anti-eNOS antibody detected by goat anti-mouse IgG linked to FITC.
Images were processed from three independent experiments and the staining was
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.

168

CHAPTER 6
CONCLUSIONS AND FUTURE DIRECTIONS
Until the late 1980s, the process of amino acid transport into mammalian cells
was attributed to a loosely-defined group of "systems" with overlapping properties.
Substrate specificity, many regulatory properties, and kinetic parameters were well-
characterized, yet the actual proteins responsible for the activities remained unknown.
Within the last decade, more than two dozen amino acid transporters have been
identified, cloned, and categorized into families according to intrinsic structural and
functional properties. Mutagenesis studies have provided information concerning
substrate binding and transport activity. Northern analyses and immunoblots have
provided evidence concerning the differential expression and regulation of members of
the same transporter family. However, very little is known regarding the "life cycle" or
the active trafficking of amino acid transporters. It is believed that amino acid
transporters follow many of the individual steps, from biosynthesis to degradation, that
have been delineated for other plasma membrane proteins. With antibodies and
molecular tags available, it has been possible to study the subcellular localization and
trafficking of specific amino acid transporters. The individual projects that have
contributed to this thesis share a common interest in documenting the cellular
localization, and when possible, understanding the functional consequences of transporter
protein distribution.
169

170
The experiments performed in Chapter 3 were designed to compare and contrast
the cellular localization of amino acid transporters within the same gene family.
Although antibodies to four of the five cloned glutamate transporters were available, only
EAAT1 and EAAT3 were detected in human fibroblasts by immunofluorescence.
Whereas both of the glutamate transporters were detected in intracellular vesicle pools,
additional immunoreactivity for EAAT3 was shown to cluster on the plasma membrane,
whereas that for EAAT1 was observed in the nuclear membrane. Both of these
observations were interesting and unexpected. Although EAAT1 and EAAT3 belong to
the same gene family, the EAAT3 clustering pattern is similar to that of the unrelated
CAT1 transporter. The validity of the EAAT1 nuclear staining remains in question;
however, it raises some intriguing possibilities. Even if the EAAT1-R antibody is
detecting a protein other than EAAT1, it is possible that a protein related to EAAT1
resides in the nucleus. More extensive immunoblotting must be performed to determine
if contamination of other membranes was contributing to the immunoreactivity of
EAAT1 in the nuclear fraction. Conclusive proof must await antibody-independent
identification of the immunoreactive protein within the nuclear membrane.
The experiments in Chapter 4 were designed to test the hypothesis that the
abnormal amino acid transport observed in LPI fibroblasts was due to aberrant trafficking
of membrane proteins, including transporters. Although the CAT1 staining appeared to
be normal, a thorough investigation of the organelles involved in the cellular trafficking
pathways detected a morphological difference between the lysosomes of the normal and
LPI fibroblasts. It is suspected that the enlarged and overly abundant LPI vacuoles are
lysosomes based on their detection by antibodies against cathepsin D and the lysosomal

171
membrane protein, lpgl20. In the future it will be important to isolate the lysosomes
from normal and LPI fibroblasts for further characterization. The sizes of the lysosomal
vesicles isolated from normal and LPI cells can be measured and compared by flow
cytometry. Additionally, measuring the pH will determine if the abnormality is linked to
a change in the acidity of the lysosomal compartment in LPI cells. There is no evidence
to suggest that the CAT1 plasma membrane transporter is contained in the abnormal LPI
vesicle population, however, we hypothesize that the vesicles may accumulate elevated
levels of cationic amino acids via the lysosomal cationic amino acid transport system
(system c). Performing transport studies that compare the influx and efflux of various
amino acids in the normal and LPI lysosomal preparations would be necessary to
investigate this hypothesis. There are several possibilities that may result in the abnormal
size of the LPI vacuoles. As mentioned above, accumulation of amino acids, other small
molecules, or fluids may result in the swelling of the vacuoles. Alternatively, the
contents of the vesicles may be normal, and the enlarged size due to accelerated
membrane fusion or reduced membrane fission. It will be important to determine which
of the above explanations, if any, applies to the abnormal LPI vesicle morphology.
Chapter 5 explores beyond the distribution of amino acid transporters, and
proposes functional consequences for the specific localization of the CAT1 amino acid
transporter in endothelial cells. This project was an extension of previous experiments in
our laboratory by Dr. Michelle Woodard, who first observed the CAT1 transporter
"patching" pattern in several cell types. Co-localization experiments strongly suggested
that CAT1 was localized to plasma membrane and/or Golgi caveolae, where it may be
directly associated with eNOS. If a complex truly exists, it would provide a mechanistic

172
explanation for the "arginine paradox," an observation that the cellular eNOS
preferentially uses extracellular arginine for NO production despite the high, even
saturating, intracellular levels of the substrate. A physical interaction between CAT1 and
eNOS would also introduce a novel role for the CAT1 transporter in signal transduction
via the production of NO.
Several future experiments will be conducted to determine if an actual interaction
exists between CAT1 and eNOS. Initial co-precipitation experiments using the CAT1
antibody to precipitate eNOS failed. Therefore, future experiments will attempt to co
precipitate eNOS using an antibody against the GFP(C3)-CAT1 fusion protein. Using a
commercially available antibody generated against the GFP tag will provide an
alternative in case the CAT1 antibody is not suitable for the immunoprecipitation
protocol. Additional future experiments will include the examination of the intracellular
co-localization of CATl/10e6 and eNOS/sialyltransferase dual-staining experiments,
however, it is unclear whether an interaction exists in this compartment. Also, the
integrity of the CATl/eNOS complex following the disruption of actin and other
cytoskeletal components will be explored further.
The data presented in this thesis begin to answer some questions regarding the cell
biology of amino acid transporters. As the molecular tools improve, and as more
transporters are identified and used to generate antibodies, a more thorough investigation
of the "life cycle" of amino acid transporters will be possible.

LITERATURE CITED
Agard, D. A. (1984) Ann. Rev. Biophys. Bioeng. 13, 191-219
Agard, D. A., Hiraoka, Y Shaw, P. and Sedat, J. W. (1989) Methods Cell Biol. 30,
353-377
Agard, D. A. and Sedat, J. W. (1983) Nature 302, 676-681
Aisaka, K., Gross, S. S., Griffith, O. W., and Levi, R. (1989) Biochem. Biophys. Res.
Commun. 163,710-717
Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J. D. (1994) in
Molecular Biology of the Cell, third edition, Garland Publishing, Inc., New York
Albritton, L. M., Kim, J. W., Tseng, L. and Cunningham, J. M. (1993) J. Virol. 67,
2091-2096
Albritton, L. M., Tseng, L., Scadden, D. and Cunningham, J. M. (1989) Cell 57, 659-666
Amara, S. G. and Kuhar, M. J. (1993) Rev. Neurosci. 16, 73-93
Amato, A., Barbour, B., Szatkowski, M. and Attwell, D. (1992) J. Physiol. 479, 371-380
Amato, A., Barbour, B., Szatkowski, M. and Attwell, D. (1994) J. Physiol. Lond. 479,
371-380
Anderson, R. G. W. (1993) Trends Cell Biol. 3, 69-72
Aris, J. P. and Blobel, G. (1988) J. Cell Biol. 107, 17-31
Amal, J.-F., Yamin, J., Dockery, S., and Harrison, D. G. (1994) Am. J. Physiol. 267,
C1381 -C1388
Arriza, J. L., Eliasof, S., Kavanaugh, M. P. and Amara, S. G. (1997) Proc. Natl. Acad.
Sci. U.S.A. 94,4155-4160
Arriza, J. L., Fairman, W. A., Wadiche, J. L, Murdoch, G. H., Kavanaugh, M. P. and
Amara, S. G. (1994) J. Neurosci. 14,5559-5569
Arriza, J. L., Kavanaugh, M. P., Fairman, W. A., Wu, Y. N., Murdoch, G. H., North, R.
A. and Amara, S. G. (1993) J. Biol. Chem. 268, 15329-15332
173

174
Baldini, N., Scotlandi, K., Serra, M., Shikita, T., Zini, N., Ognibene, A., Santi, S.,
Ferracini, R. and Maraldi, N. (1995) Eur. J. Cell Biol. 68, 226-239
Bar-Peled, O., Ben-Hur, H., Biegon, A., Groner, Y., Dewhurst, S., Furuta, A. and
Rothstein, J. D. (\997) J. of Neurochem. 69,2571-2579
Baydoun, A. R., Emery, P. W., Pearson, J. D., and Mann, G. E. (1990) Biochem.
Biophys. Res. Commun. 173, 940-948
Bertrn, J., Magagnin, S., Werner, A., Markovich, D., Biber, J., Testar, X., Zorzano,
Kuhn, L. C., Palacin, M. and Murer, H. (1992) Proc. Nat. Acad Sci. U.S.A. 89,
5606-5610
Bhat, G. B., and Block, E. R. (1992) Am. J. Respir. Cell. Mol. Biol. 6, 633-638
Bhat,G. B. and Block, E. R. (1990) Am. J. Respir. Cell. Mol. Biol. 3, 363-367
Bliss, T. V. and Collingridge, G. L. (1993) Nature (Lond.) 361, 31-39
Block, E. R., Herrera, H., and Couch, M. (1995) Am. J. Physiol. 269, L574-L580
Block, E. R Patel, J. M. and Edwards, D. (1989) Cell Physiol. 26, C223-C231
Boger, R1 H., Bode-Boger, S. M., Thiele, W., Junker, W., Alexander, K. and Frolich, J.
C. (1997) Circulation 95, 2068-2074
Boulikas, Teni (1996) J. Cell Biochem. 60,61-82
Bouvier, M., Szatkowski, M., Amato, A. and Attwell, D. (1992) Nature 360, 471-474
Bu, G. and Schwartz, A. L. (1994) in Cell Biology: A Laboratory Handbook pp. 199-
200, Academic Press, Inc.
Buchanan, J. A., Rosenblatt, D. S. and Scriver, C. R. (1985),4wj. NY. Acad. Sci. 256,
401-402
Bussolati, O., Laris, P. C., Nucci, F. A., Dall'Asta, V., Longo, N., Guidotti, G. G. and
Gazzola, G. C. (1987) Am. J. Physiol. 253, C1-C7
Bussolati, O., Sala, R., Astorri, A., Rotoli, B. M., DallAsta, V., and Gazzola, G. C.
(1993) Am. J. Physiol. 265, C1006-C1014
Calonge, M. J., Gasparini, P., Chillaron, J., Chilln, M., Gallucci, M., Rousaud, F.,
Zelante, L., Testar, X., Dallapiccola, B., Di Silverio, F., Barcelo, P., Zorzano, A.,
Nunes, V. and Palacin, M. (1994) Nat. Genet. 6, 420-425
Carlsson, M. and Carlsson, A., (1990) Trends Neurosci. 13, 272-276

175
Chen, H., Hughes, D. D., Chan, T., Sedat, J. W., and Agard, D. A. (1996) J. Structural
Biol. 116,56-60
Chen, H., Sedat, J. W., and Agard, D. A. (1989) in Handbook of Biological Confocal
Microscopy pp. 127-135,1.M.R. Press, Madison, WI
Cheng, S. H., Gregory, R. J., Marshall, J., Paul, S., Souza, D. W., White, G. A.,
O'Riordan, C. R. and Smith, A. E. (1990) Cell 63, 827-834
Chiles, T. C., O'Brien, T. W. and Kilberg M. S. (1987) Anal. Biochem. 163, 136-142
Cho, K. 0., Hunt, C. A. and Kennedy, M. B. (1992) Neuron 9, 929-942
Chou, H.-F., Vadgama, J. and Jonas, A. J. (1992) Chem. Med. Metabol. Biol. 48, 179-
193
Chun, M., Liyanage, U. K., Lisanti, M. P., and Lodish, H. F. (1994) Proc. Natl. Acad.
Sci, U. S. A. 91, 11728-11732
Clements, J. D Lester, R. A. J., Tong, G., Jahr, C. E. and Westbrook, (1992) Science
258, 1498-1501
Closs, E. I. (1996) Amino Acids 11, 193-208
Closs, E. I., Albritton, L. M. Kim, J. W. and Cunningham, J. M. (1993) J. Biol. Chem.
268, 7538-7544
Colley, N. J Baker, E. K Stamnes, M. A. and Zuker, C. S. (1991) Cell 67, 255-263
Conrad, P. A., Smart, E. J., Ying, Y-S., Anderson, R. G. W., and Bloom, G. (1995) J.
Cell Biol. 131,1421-1433
Conradt, M. and Stoffel, W. (1997) J. ofNeurochem. 68, 1244-1251
Conradt, M Storck, T and Stoffel, W. (1995) Eur. J. Biochem. 229, 682-687
Cooke, J. P Andn, N. A., Girerd, X. J., Hirsch, A. T., and Creager, M. A. (1991)
Circulation 83, 1057-1062
Cormack, B. P Valdivia, R. and Falkow, S. (1996) Gene 173, 33-38
Cowbum, R., Hardy, J., Roberts, P. and Briggs, R. (1988) Neurosci. Lett. 86, 109-113
Creager, M. A., Gallagher, S. J., Girerd, X. J., Coleman, S. M., Dzau, V. J., and Cooke, J.
P. (1992) J. Clin. Invest. 90, 1248-1253
Dall'Asta, V., Gazzola, G. C., Franchi-Gazzola, R., Bussolati, O., Longo, N. and Guidotti,
G. G. (1983) J. Biol. Chem. 258,6371-6379

176
Davies, J. P., Cotter, P. D. and Ioannou, Y. A. (1997) Genomics 41, 131-134
Davis, K. E., Straff, D. J., Weinstein, E. A., Bannerman, P. G., Corrale, D. M.,
Rothstein, J. D. and Robinson, M. B. (1998) J. Neurosci. 18, 2475-2485
Davis, L. I. and Blobel, G. (1986) Cell 45, 699-709
De Silva, D. M Askwith, C. C. and Kaplan, J. (1996) Physiol Rev. 76, 31-44
de Weerd, W. F. and Leeb-Lundberg, L. M. (1997) J. Biol. Chem. 272, 17858-17866
Deves, R., Chavez, P. and Boyd, C. A., R. (1992) J. Physiol., Lond. 454, 491-501
Drumm, M. L., Wilkinson, D. J., Smit, L. S., Worrell, R. T., Strong, T. V., Frizzell, R.
A., Dawson, D. C. and Collins, F. S. (1991) Science 254, 1797-1799
Eddahibi, S., Adnot, S., Carville, C., Blouquit, Y., and Raffestin, B. (1992) Am. J.
Physiol. 263, L194-L200
Fafournoux, P., Dudenhausen, E. E., and Kilberg, M. S. (1989) J. Biol. Chem. 264,
4805-4811
Fairman, W. A., Vandenberg, R. J., Arriza, J. L., Kavanaugh, M. P. and Amara, S. G.
(1995) Nature 375, 599-603
Feron, O., Belhassen, L., Kobzik, L., Smith, T. W., Kelly, R. A., and Michel, T. (1996)
J. Biol. Chem. 271,22810-22814
Feron, 0., Smith, T. W Michel, T. and Kelly, R. A. (1997) J. Biol. Chem. 272, 17744-
17748
Forstermann, U., Closs, E. I., Pollock, J. S., Nakane, M., Schwarz, P., Gath, I., and
Kleinert, H. (1994) Hypertension 23, 1121-1131
Garcia-Cardena, G., Fan, R Stem, D. F., Liu, J., and Sessa, W. C. (1996) J. Biol. Chem.
271, 27237-27240
Garcia-Cardena, G., Oh, P., Liu, J., Schnitzer, J. E., and Sessa W. C. (1996) Proc. Natl.
Acad. Sci. USA 93, 6448-6553
Ginsberg, S. D., Martin, L. J. and Rothstein, J. D. (1995) J. ofNeurochem. 65, 2800-
2803

177
Gold, M. E., Bush, P. A., and Ignarro, L J. (1989) Biochem. Biophys. Res.Comunm. 164,
714-721
Greene, B., Pacitti, A. J., Souba, W. W. (1993) Am. J. Physiol. 264, L351-L356
Hanson, S. H., Sandvig, K. and van Deurs, B. (1993) J. Cell Biol. 123, 89-97
Harrison, D. G. (1996) Cardiol. Clin. 14, 1-15
Harrison, D. G. (1997) J. Clin. Invest. 100,2153-2157
Haugeto, O., Ullensvang, K., Levy, L., M., Chaudhry, F. A., Honor, T., Nielsen, M.,
Lehre, K. P. and Danbolt, N. C. (1996) J. Biol. Chem. 271, 27715-27722
Hecker, M., Sessa, W. C., Harris, H. J., Anggard, E. E., and Vane, J. R. (1990) Proc.
Natl. Acad. Sci. USA 87, 8612-8616
Hemler, M. E., Sanchez-Madrid, F., Flotte, T. J., Krensky, A. M., Burakoff, S. J., Bhan,
A. K., Springer, T. A. and Strominger, J. L. (1984) J. Immunol. 132, 3011-3018
Hershko, A. and Ciechanover, A. (1992) Annu. Rev. Biochem. 61, 761-807
Hiraoka, Y., Sedat, J. W., and Agard, D. A. (1987) Science 238, 36-41
Hiraoka, Y., Swedlow, J. R., Paddy, M. R., Agard, D. A. and Sedat, J. W. (1991) Sem.
Cell Biol. 2, 153-165
Hirata, K., Miki, N., Kuroda, Y., Sakoda, T., Kawashima, S. and Yokoyama, M. (1995)
Circ. Res. 76, 958-962
Hosokawa, H., Sawamura, T., Kobayashi, S., Ninomiya, H., Miwa, S. and Masaki, T.
(1997) J. Biol. Chem. 272,8717-8722
Ito, K. and Groudine, M. (1997) J. Biol. Chem. 272, 26780-26786
Ivessa, N. E., De Lemos-Chiarandini, C., Gravotta, D., Sabatini, D. D. and Kreibich, G.
(1995) J. Biol. Chem. 270,25960-25967
Jones, K. L., Stewart, R. M., Fowler, M., Fukuda, M. and Holcombe, R. F. (1992) Clin.
Immunol. Immunopathol. 65, 219-226
Juliano, R. L. and Ling, V. (1976) Biochim. Biophys. Acta 455, 152-162
Kakuda, D. K., Finley, K. D., Dionne, V. E. and MacLeod, C. L. (1993) Transgene 1,
91-101

178
Kanai, Y. and Hediger, M. A. (1992) Nature 360, 467-471
Kanai, Y., Nussberger, S., Romero, M. F., Boron, W. F., Hebert, S. C. and Hediger, M.
A., (1995) J. Biol. Chem. 270, 16561-16568
Kanai, Y., Stelzner, M. G., Lee, W.-S., Wells, R. G., Brown, D. and Hediger, M. A.
(1992) Am. J. Physiol. 263, F1087-F1093
Kekuda, D. K., Prasad, P. D., Fei, Y.-J., Torres-Zamorano, V., Sinha, S., Yang-Feng, T.
L., Leibach, F. H. and Ganapathy, V. (1996) J. Biol. Chem. 271, 18657-18661
Kawakami, H., Tanaka, K., Nakayama, T., Inoue, K. and Nakamura, S. (1994) Biochem.
Biophys. Res. Commun. 199, 171-176
Kerem, B.-S., Rommens, J. M., Buchanan, J. A., Markiewicz, D., Cox, T. K.,
Chakravarti, A., Buchwald, M. and Tsui, L.-C. (1989) Science 245, 1073-1080
Kilberg, M. S., Stevens, B. R and Novak, D. (1993) Annu. Rev. Nutr. 13, 137-165
Kim, J. W., Class, E. I., Albritton, L. M., and Cunningham, J. M. (1991) Nature 352,
725-728
Klockner, U Storck, T., Conradt, M. and Stoffel, W. (1993) J. Biol. Chem. 268, 14594-
14596
Komfeld, S. and Mellman, I. (1989) Annu. Rev. Cell Biol. 5, 483-525
Kurz, S. and Harrison, D. G. (1997) J. Clin. Invest. 99, 369-370
Laemmli, U. K. (1970) Nature 227, 680-685
Laine, R. O. and Esser, A. F. (1989) J. Immunol. 143, 553-557
Li, S., Mallory, M., Alford, M., Tanaka, S. and Masliah, E. (1997) J. Neuropathol. Exp.
Neurol. 56,901-911
Liao, K. and Lane M. D. (1995) Biochem. Biophys. Res. Commun. 208, 1008-1015
Lin, C. L., Bristol, L. A., Jin, L., Dykes-Hoberg, M., Crawford, T., Clawson, L. and
Rothstein, J. D. (1998) Neuron 20, 589-602
Lisanti, M. P., Scherer, P. E., Tang, Z., and Sargiacomo, M. (1994) Trends Cell Biol. 4,
231-235
Liu, J., Garcia-Cardena, G., and Sessa, W. C. (1996) Biochemistry 35, 13277-13281
Ljungdahl, P. O., Gimeno, C. J., Styles, C. A. and Fink, G. R. (1992) Cell 71, 463-478

179
MacLeod, C. L., Finley, K. D., Kakuda, D.K. (1994) J. Exp. Biol. 196, 109-21
MacLeod, C. L., Finley, K., Kakuda, D., Kozak, C. A. and Wilkinson, M. F. (1990) Mol.
Cell. Biol. 10, 3663-3674
Makowske, M. and Christensen, H. N. (1989) J. Biol. Chem. 257, 5663-5670
Malandro, M. S. and Kilberg, M. S. (1996) Annu. Rev. Biochem. 65, 305-336
Mancini, G. M. S., Beerens, E. M. T., Aula, P. P. and Verheijen, F. W. (1991) J Clin.
Invest. 87, 1329-1335
Manfras, B. J., Rudert, W. A., Trueco, M. and Boehm, B. O. (1994) Biochim. Biophys.
Acta, 1195, 185-188
Matthews, J. C., Beveridge, M. J., Malandro, M. S., Rothstein, J. D., Campbell-
Thompson, M., Verlander, J. W., Kilberg, M. S. and Novak, D. A. (1998) Am. J.
Physiol. 274, C603-C614
Mattson, M. P., Lee, R. E., Adams, M. E., Gutherie, P. B. and Kater, S. B. (1988) 1
Neuron 1, 865-876
McDonald, K. K., Rouhani, R., Handlogten, M.E., Block, E. R., Griffith, O. W., Allison,
R. D and Kilberg, M. S. (1997) Biochim. Biophys. Acta 1324, 133-141
McDonald, K. K Zharikov, S Block, E. R and Kilberg, M. S. (1997) J. Biol. Chem.
272,31213-31216
Michel, J., Feron, 0., Sacks, D. and Michel, T. (1997) J. Biol. Chem. 272, 15583-15586
Mitchell, J. A.,Flecker, M., Anggard, E. E., and Vane, J. R. (1990) Eur. J. Pharmacol.
182, 573-576
Moneada, S. and Higgs, A. (1993) N. Engl. J. Med. 329, 2002-2012
Moneada, S., Palmer, R. M., and Higgs, E. A. (1991) Pharmacol. Rev. 43, 109-142
Morin, A. M. and Stanboli, A. (1993) J. Neurosci. Res. 36, 272-279
Mugge, A. and Harrison, D. G. (1991) Blood Vessels 28, 354-357
Nathan, C. and Xie, Q. (1994) Cell 78, 915-918
Nicholls, D. G. (1993) Eur. J. Biochem. 212, 613-631
Nishida, K., Harrison, D. G., Navas, J. P., Fisher, A. A., Dockery, S. P., Uematsu, M.,
Nerem, R. M., Alexander, R. W. and Murphy, T. J. (1992) J. Clin. Invest. 90,
2092-2096

180
Nuoffer, C. and Balch, W. E. (1994)Annu. Rev. Biochem. 63, 949-990
O'Brien, H. A., Young, H. N., Pobey, J. M. and Furness, J. V. (1995) Histochemistry
103, 221-225
Okamoto, T., Schlegel, A., Scherer, P. E. and Lisanti, M. P. (1998) J. Biol. Chem. 273,
5419-5422
Otori, Y., Shimada, S., Tanaka, K., Ishimoto, I., Tao, Y. and Tohyama, M. (1994) Mol.
Brain Res. 27, 310-314
Oxender, D. L. and Christensen, H. N. (1963) J. Biol. Chem. 238, 3686-3699
Paddy, M. R., Belmont, A. S., Saumweber, H., Agard, D. A. and Sedat, J. W. (1990)
Cell 62,89-106
Palacin, M. (1994) J. Exp. Biol. 196, 123-137
Palmer, R. M. and Moneada, S. (1989) Biochem. Biophys. Res. Commun. 158, 348-352
Palmer, R. M., Rees, D. D., Ashton, D. S., and Moneada, S. (1988) Biochem. Biophys.
Res. Commun. 153, 1251-1256
Parton, R. G. (1996) Curr. Opin. Cell. Biol. 8, 542 -548
Pearce, I. A., Cambray-Deakin, M. A. and Burgoyne, R. D. (1987) FEBS Lett. 223, 143-
147
Peghini, P Janzen, J. and Stoffel, W. (1997) EMBO 16, 3822-3832
Pelham, G. R. (1991) Cell 67, 449-451
Perou, C. M., Leslie, J. D., Green W., Li, L., Ward, D. M. and Kaplan, J. (1997) J. Biol.
Chem. 272,29790-29794
Pind, S Riordan, J. R. and Williams, D. B. (1994) J. Biol. Chem. 269, 12784-12788
Pines, G., Danbolt, N. C., Bjoras, M., Zhang, Y, Bendahan, A., Eide, L., Koepsell, H.,
Storm-Mathisen, J., Seeberg, E. and Kanner, B. I. (1992) Nature 360, 464-467
Pines, G., Zhang, Y. and Kanner, B. I. (1995) J. Biol. Chem. 270, 17093-17097
Pisoni, R. L., Thoene, J. G. and Christensen, H. N. (1985) J. Biol. Chem. 260,4791-
4798
Podlecki, D. A., Smith, R. M., Kao, M., Tsai, P., Huecksteadt, T., Brandenburg, D.,
Lasher, R. S., Jarett, L. and Olefsky, J. M. (1987) J. Biol. Chem. 262, 3362-
3368

181
Pollock, J. S., Forstermann, U., Mitchell, J. A., Warner, T. D., Schmidt, H. H., Nakane,
M and Murad, F. (1991) Proc. Natl. Acad. Sci. USA 88, 10480-10484
Rajantie, J., Simell, O. and Perheentupa, J. (1981) J Clin. Invest. 67, 1078-1082
Rakowicz-Szulczynska, E. M., Linnenbach, A. J. and Koprowski, H. (1989) Mol.
Carcinog. 2, 47-58
Ravichandran, L. V. and Johns, R. A. (1995) FEES Lett. 374, 295-298
Robbins, E. and Marcus, P. I. (1963) J. Cell Biol. 18, 237-250
Robinson, M. B., Hunter-Ensor, M. and Sinor, J. (1991) Brain Res. 544, 196-201
Rossitch, E., Jr., Alexander, E., Black, P. M., and Cooke, J. P. (1991) J. Clin. Invest. 87,
1295-1299
Rothstein, J. D., Dykes-Hoberg, M., Pardo, C. A., Bristol, L. A., Jin, L., Kunci, R. W.,
Kanai, Y., Flediger, M. A., Wang, Y., Schielke, J. P. and Welty, D. F. (1996)
Neuron 16, 675-686
Rothstein, J. D., Martin, L. and Kunci, R. W. (1992) N. Engl. J. Med. 326, 1464-1468
Rothstein, J. D., Van Kammen, M., Levey, A. I., Martin, L. and Kunci, R. W. (1995)
Ann. Neurol. 38, 73-84
Schneider, J. A. and Schulman, J. D. (1982) in The Metabolic Basis of Inherited Disease
(Scriver, C. R., Beaudet, A. L., Sly, W. S. and Valle, D., ed) pp. 1844-1866,
McGraw-Hill, New York
Schnitzer, J. E Oh, P., Pinney, E and Allard, J. (1994) J. Cell Biol. 127, 1217-1232
Scott, H. L., Tannenberg, A. E. G. and Dodd, P. R. (1995) J. Neurochem. 64, 2193-2202
Segal, S. and Thier, S. O. (1989) in The Metabolic Basis of Inherited Disease (Scriver,
C. R., Beaudet, A. L., Sly, W. S. and Valle, D., ed) pp. 2479-2496, McGraw-Hill,
New York
Sessa, W. C., Garcia-Cardena, G., Liu, J., Keh, A., Pollock, J. S., Bradley, J., Thiru, S.,
Braverman, I. M. and Desai, K. M. (1995) J. Biol. Chem. 270, 17641-17644
Shafquat, S., Tamarappoo, B. K., Kilberg, M. S., Puranam, R. S., McNamara, J. O.,
Guadano-Ferraz, A. and Fremeau, R. T. (1993) J. Biol. Chem. 268, 15351-15355
Shaul, P. W., Smart, E. J., Robinson, L. J., German, Z., Yuhanna, I. S., Ying, Y.,
Anderson, R. G., and Michel, T. (1996) Biol. Chem. 271, 6518-6522

182
Shayakul, C., Kanai, Y., Lee, W.-S., Brown, D., Rothstein, J. D. and Hediger, M. A.
(1998) Am. J. Physiol. 273, F1023-F1029
Shupliakov, O., Brodin, L., Cullheim, S., Ottersen, O. P. and Storm-Mathisen, J. (1992)
J. Neurosci. 12, 3789-3803
Silbemagl, S. (1983) Pflugers Arch. 396,218-224
Simell, O. (1989) in The Metabolic Basis of Inherited Disease (Scriver, C. R., Beaudet,
A. L., Sly, W. S. and Valle, D., ed) pp. 2497-2513, McGraw-Hill, New York
Simell, O., Perheentupa, J., Rapla, J., Visakorpi, J. K. and Eskelin, L.- E. (1975) Am. J.
Med. 59, 229-240
Simionescu, M., Simionescu, N. (1987) in Pulmonary Endothelium in Health and
Disease (US Ryan, ed) pp. 35-62, Marcel Dekker, Inc., New York
Smith, D. W., Scriver, C. R., Tenenhouse, H. S. and Simell, O. (1987) Proc. Natl. Acad.
Sci. 84, 7711-7715
Smith, D. W., Scriver, C. R. and Simell, O. (1988) Hum. Genet. 80, 395-396
Solski, P. A., Quilliam, L. A., Coats, S. G., Der, C. J. and Buss, J. E. (1995) Meth.
Enzym. 250, 435-444
Song, K. S., Li, S., Okamoto, T., Quilliam, L. A., Sargiacomo, M. and Lisanti, M. P.
(1996) J. Biol. Chem. 271, 9690-9697
Stachowiak, M. K., Maher, P. A., Joy, A., Mordechai, E. and Stachowiak, E. (1996) Mol.
Cell Biol. 7,1299-1317
Steams, T. (1995) Curr. Biol. 5,262-264
Storck, T., Schulte, S., Hofmann, K. and Stoffel, W. (1992) Proc. Natl. Acad. Sci. U.S.A.
89, 10955-10959
Storm-Mathisen, J., Danbolt, N. C., Rothe, F., Torp, R., Zhang, N., Aas, J. E., Kanner, B.
L, Langmoen, I. and Ottersen, O. P. (1992) Prog. Brain Res. 94, 225-241
Szatkowski, M. and Attwell, D. (1994) Trends Neurosci. 17, 359-365
Tamarappoo, B. K., Handlogten, M. E., Laine, R. O., Serrano, M. A., Dugan, J., and
Kilberg, M. S. (1992) J. Biol. Chem. 267, 2370-2374
Tamarappoo, B. K., McDonald, K. K. and Kilberg, M. S. (1996) Biochim. Biophys. Acta
1219, 131-136

183
Tanaka, K., Watase, K., Manabe, T., Yamada, K., Watanabe, M., Takahashi, K., Iwama.
H., Nishikawa, T., Ichihara, N., Kikuchi, T., Okuyama, S., Kawashima, N., Hori,
S. and Takimoto, M. (1997) Science 276, 1699-1702
Taylor, P. D.and Poston, L. (1994) Br. J. Pharmacol. 113, 801-808
Teitel, J. M. (1986) J. Biol. Chem. 128,329-336
Utsunomiya-Tate, N., Endou, H. and Kanai, Y. (1996) J. Biol. Chem. 271, 14883-14890
Vadgama, J. V. and Christensen, H. N. (1984) J. Biol. Chem. 259, 3648-3652
Van Aelst, L., Barr, M., Marcus, S., Polverino, A. and Wigler, M. (1993) 90 Proc. Natl.
Acad. Sel, U. S. A.. 90, 6213-6217
Van Slyke, D. D. and Meyer, G. M. (1913) J. Biol. Chem. 16, 197
Van Winkle, L. J. (1988) Biochim. Biophys. Acta 947, 173-208
Van Winkle, L. J., Campione, A. L. and Gorman, M. L. (1988) J. Biol. Chem. 263,
3150-3163
Van Winkle, L. J., Christensen, H. N. and Campione, A. L. (1985) J. Biol. Chem. 260,
12118-12123
Velaz-Faircloth, M., McGraw, T. S., Malandro, M. S., Fremeau, R. T., Kilberg, M. S.,
and Anderson, K. J. (1996) Am. J. Physiol. 270, 67-75
Wahle, S. and Stoffel, W. (1996) J. Cell Biol. 135, 1867-1877
Wang, H., Kavanaugh, M. P., North, R. A., and Rabat, D. (1991) Nature 352, 729-731
Wang, Y. and Tate, S. S. (1995) FEBSLett. 368, 389-392
Ward, D. M Leslie, J. D. and Kaplan, J. (1997) J. Cell Biol. 139, 665-673
Watts, C. and Marsh, M. (1992) J. Cell Sci. 103, 1-8
Wells, R. G., Lee, W., Kanai, Y., Leiden, J. M. and Hediger, M. A. (1992) J. Biol. Chem.
267, 15285-15288
White, M. F. (1985) Biochim. Biophys. Acta 822, 355-374
White, M. F. and Christensen, H. N. (1983) J. Biol. Chem. 258, 8028-8038
White, M. F., Gazzola, G. C and Christensen, H. N. (1992) J. Biol. Chem. 257, 4443-
4449

184
Woodard, M. H., Dunn, W. A., Laine, R. O., Malandro, M., McMahon, R., Simell, O.,
Block, E. R., and Kilberg, M. S., (1994) Am. J. Physiol. 266, E817-E824
Wu, J. Y., Robinson, D., Kung, H., and Hatzoglou, M. (1994) J. Virol., 68, 1615-1623
Yoshimoto, T., Yoshimoto, E. and Meruelo, D. (1991) Virology 185, 1-17
Zembowicz, A., Tang, J. and Wu, K. K. (1995) J. Biol. Chem. 270, 17006-17010
Zerangue, N. and Kavanaugh, M. P. (1996) Nature 383, 634-637
Zhang, Y., Pines, G. and Kanner, B. I. (1994) J. Biol. Chem. 269, 19573-19577
Zharikov, S. I. and Block, E. R. (1997) Biochim. Biophys. Acta, in revision

BIOGRAPHICAL SKETCH
Kelly Kristin McDonald was bom to Maurice and Patricia McDonald on August
1, 1971, in Nashville, Tennessee. After graduating from John Overton High School, in
Nashville, Tennessee, she attended Western Kentucky University where she studied
theatre and dance as a Performing Arts major for three years. In 1992, Kelly transferred
to the University of Florida and earned a B. S. in Biochemistry. She began graduate
studies in the Department of Biochemistry and Molecular Biology at the University of
Florida in January 1994. Following graduation, Kelly will begin work on a Master's in
Scientific Journalism and Mass Communications at the University of Florida. In
addition, she will continue to conduct research with Dr. Edward Block at the University
of Florida College of Medicine.
185

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Michael S. Kilberg, Chair (J
Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
i
Brian D. Cain
Associate Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor of Philosophy.
William A. Dunn
Associate Professor of Anatomy and
Cell Biology
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
"Susan C. Frost
Associate Professor of Biochemistry
and Molecular Biology
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
phy/' \
Peter M. McGuire
Assistant Professor of Biochemistry
and Molecular Biology

This dissertation was submitted to the Graduate Faculty of the College of
Medicine and to the Graduate School and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
August, 1998
Dean, College of Medicine
Dean, Graduate School

iSi



5-9. Immunofluorescent staining of PAEC transfected with the
CATMUT1 palmitoylation mutant 164
5-10. Immunofluorescent staining of PAEC transfected with the
CATMUT3 palmitoylation mutant 166
5-11. eNOS staining of PAEC treated with varying concentrations of
extracellular L-arginine 168
vm


LIST OF FIGURES
Figure page
3-1. Extracellular staining of human fibroblasts with EAAT3 antibody 53
3-2. Intracellular staining of human fibroblasts with EAAT3 antibody
and co-localization with organelle-specific antibodies 55
3-3. Nuclear staining of human fibroblasts with EAAT1 -R antibody and
co-localization with nucleus-specific antibodies 57
3-4. Intracellular staining of human fibroblasts with EAAT1 -S and
EAAT1-C antibodies 59
3-5. Intracellular staining of Hela cells with EAAT1-R and EAAT1-S
antibodies 61
3-6. Immunoblot analysis of EAAT1 in the nuclear and intracellular
membrane fractions from human fibroblasts 62
3-7. Expression of GFP and GFP-EAAT1 fusion proteins in human
fibroblasts 64
3-8. EAAT1 immunofluorescent staining of human fibroblasts
transfected with EAAT1-GFP(N3) 66
3-9. EAAT1 immunofluorescent staining of PAEC transfected with
EAAT1-GFP(N3) 68
4-1. Morphology of normal and LPI fibroblasts by light microscopy 97
4-2. Morphology of normal and LPI fibroblasts by electron microscopy 99
4-3. Intracellular staining of normal and LPI human fibroblasts with the
CAT1 antibody 101
4-4. Expression of GFP and the GFP(C3)-CAT1 fusion protein in
normal human fibroblasts 103
4-5. Expression of the GFP(C3)-CAT1 fusion protein in normal and
LPI human fibroblasts 105
vi


Figure 5-10. Immunofluorescent staining of PAEC transfected with the CATMUT3
palmitoylation mutant. PAEC were transfected for 3 h with the CATMUT3 fusion
protein that has serine substitutions for cysteine at residues 3, 20, and 30, and an alanine
substitution for glycine at residue 2, and was generated according to the mutagenesis and
transfection protocols described in the Methods Chapter. Transfection of the PAEC was
performed according to the procedures described in the Methods Chapter. Following 24-
48 h of expression, cells were fixed with -20C MeOH and stained with antibodies
against caveolin (A) and eNOS (B). Both the caveolin and eNOS antibodies were
detected with a goat anti-rabbit IgG conjugated to Texas Red and visualized by
deconvolution microscopy. Images were processed from three independent experiments
and the staining was determined to be reproducible. The data shown represent analysis of
1.0 pm sections through the cells.


27
to check for reproducibility, then cells from a different patient were used to confirm that
the results are not unique to a single individual with the disease.


Figure 4-11. Intracellular staining of normal and LPI fibroblasts with an antibody against
a lysosomal membrane protein. According to the protocol described in the Methods
Chapter, normal (A) and LPI (B) fibroblasts were fixed with -20C MeOH and subjected
to immunohistochemistry with the anti-lpgl20 antibody specific for lysosomal
membranes. The anti-lpgl20 primary antibody was visualized with a goat anti-rabbit IgG
conjugated to FITC. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells.


157


47
fusion proteins. Although some perinuclear fluorescence was observed, probably Golgi
localization, the pattern was clearly distinct from the nuclear membrane labeling observed
with the EAAT1-R antibody.
Co-localization of EAAT1-GFPN3) and the endogenous EAAT1 in human
fibroblasts. The fluorescence pattern observed in the EAATl-GFP(N3)-transfected
fibroblasts was consistent with the pattern of staining by the EAAT1-S antibody, that is
fluorescence was associated with a population of small vesicles scattered throughout the
cytoplasm. However, to test the specificity of both the EAAT1-S and EAAT1-R
antibodies for the exogenous transporter, human fibroblasts were transfected with the
EAAT1-GFP(N3) fusion protein, fixed with -20C MeOH, and stained with either the
EAAT1-R (Figure 3-8 A) or EAAT1-S (Figure 3-8 B) antibody. The EAAT1-S and
EAAT1-R primary antibodies were detected using a 1:200 dilution of goat anti-rabbit IgG
conjugated to Texas Red. A significant amount of co-localization was observed with the
EAAT1-GFP(N3) fusion protein and EAAT1-S antibody, however, little overlap was
detected when the EAAT1-R antibody was used. In the fibroblasts stained with EAAT1-
S there were separate pools of Texas Red-labeled vesicles that did not overlap with
EAAT1-GFP(N3). This could be due to the EAAT1-S antibody recognizing endogenous
EAAT1 which, for unknown reasons, was localized in a compartment lacking the
expressed fusion protein. Conversely, there was also a population of vesicles containing
EAAT1-GFP(N3) that were not stained with EAAT1-R. This latter result can be
explained by two vesicle populations or by the fact that the antibody-antigen binding
efficiency is less than 100% for the immunofluorescence assay. These results strongly


139
Table 5-2
Immunodepletion of CATI-mediated Arginine Transport
Activity by anti-eNOS Antibody
Immunoprecipitation
Prior to Reconstitution
Transport Velocity
pmol mg-1 protein 3
Percent of Control
min'
Arsinine Transport
No antibody
3511 50

Mouse IgG
4084 + 44
100
Anti-eNOS IgG
1113 17*
27
Glutamine Transport
Mouse IgG
4026176
100
Anti-eNOS IgG
4368 706
109
* P < 0.001 versus no antibody
or non-immune mouse IgG, i
n = 3 independent
experiments
Table 5-2. After solubilization of CAT1 transport activity (Fafournoux et al., 1989), the
protein fraction was subjected to immunodepletion (Tamarappoo et al., 1992), and then
incubated with either control mouse IgG or anti-eNOS IgG bound to anti-mouse IgG
immobilized on agarose beads. After pelleting the beads, proteins remaining in the
supernatant were reconstituted into proteoliposomes to test for immunodepletion of
saturable, Na+-independent arginine transport activity (Zharikov and Block, 1997). As a
control for transporter specificity by immunodepletion, the reconstituted proteoliposomes
also were assayed for Na+-dependent 50 pM glutamine transport to demonstrate that not
all amino acid transporters were precipitated by the eNOS antibody.
To establish that the anti-eNOS did not result in immunodepletion of arginine transport in
a non-specific manner, glutamine transport was monitored in the reconstituted
proteoliposomes after immunoprecipitation with non-immune and anti-eNOS IgG. No
immunodepletion of glutamine transport was observed. These data suggest that a
functional association exists between the CAT1 arginine transporter and membrane-
bound eNOS in PAEC.
Immunoprecipitation of eNOS using anti-CATl antibody. As mentioned
previously, several laboratories have documented the co-immunoprecipitation of eNOS
and caveolin in endothelial cells (Garcia-Cardena et al., 1996). To provide additional


78
phosphatase substrate at a wavelength of 405 nm. By the sixth bleed, one of the human
CAT1 antibodies showed a 2.5-fold increase in absorbance over the pre-immune at a
dilution of 1:50. Although a series of immunohistochemistry experiments were
conducted using this anti-human CAT1 antibody, the background staining was high and
the corresponding peptide failed to completely inhibit staining. Therefore, all of the data
presented in this chapter was generated using the polyclonal murine CAT1 antibody. The
conditions for the CAT1 antibody as well as the other antibodies used in this chapter are
summarized in Table 4-1.
Table 4-1
Antibodies for Immunofluorescence Studies
Name Host Source Dilution
CAT1
rabbit
EAAT1
rabbit
EAAT3
rabbit
GLUT1
rabbit
414
mouse
D77
mouse
P-integrin
mouse
caveolin
rabbit
tubulin
mouse
KDEL
mouse
BiP
rabbit
PDI
rabbit
sialyltransferase
rabbit
Dr. Michael Kilberg,
Univ. of Florida
1:25
Dr. Jeffrey Rothstein.
John Hopkins
1:50
Dr. Michael Kilberg.
Univ. of Florida
1:200
Dr. Susan Frost.
Univ. of Florida
1:50
Dr. John Aris,
Univ. of Florida
1:10
Dr. John Aris.
Univ. of Florida
1:50
Dr. Martin Hemler,
Dana-Farber Cancer Inst.
1:100
Transduction Laboratories
Lexington, K.Y
1:200
Sigma.
St. Louis, MO
1:100
Dr. David Vaux.
EMBL
1:100
Dr. Susan Frost.
Univ. of Florida
1:50
Dr. Tom Wileman,
Dana-Farber Cancer Inst.
1:25
Dr. William Dunn.
Univ. of Florida
1:50


30
(Zhang et al., 1994). Conradt and Stoffel performed similar mutagenesis studies using
the EAAT1 cDNA (Conradt and Stoffel. 1995). When they mutated the conserved
arginine 122. arginine 280. arginine 479. and tyrosine 405, they lost glutamate transport
activity with the tyrosine 405 and arginine 479 mutants. The arginine 122 and arginine
280 mutants appeared to increase the K, of EAAT1 for aspartate, but had no effect on the
intrinsic properties or kinetics of glutamate transport. They proposed from their studies
that the hydroxyl group on tyrosine 405 and the positive charge on arginine 479 may
contribute to the binding of the acidic glutamate substrate.
Although most of the research involving the glutamate transporters has been
confined to the brain. L-glutamate is crucial to several biochemical pathways of
peripheral tissues as well (i.e., ammonia detoxification and gluconeogenesis). Several
laboratories have independently shown, by mRNA and protein analysis, that tissues other
than the brain express one or more of the transporter isoforms. EAAT3 is the most
ubiquitous of the glutamate transporters and is detected in kidney, small intestine, liver,
heart, lung, skeletal muscle, and placenta (Kanai and Hediger, 1992; Matthews et al.,
1998). Glutamate and asparate are almost completely reabsorbed from the glomerular
filtrate by the abundantly expressed EAAT3 transporter in the renal tubules (Silbemagl,
1983; Shayakul et al., 1998). EAAT1 is expressed in heart, lung, skeletal muscle, retinal
glia, and placenta (Kawakami et al., 1994; Arriza et al., 1994). and both EAAT2 and
EAAT4 have also been detected in placental tissue (Matthews et al., 1998). Preliminary
data from our laboratory (Tessmann, unpublished results) also suggest that EAAT1,
EAAT2, and EAAT3 are expressed in human fibroblasts.


37
NRDVEMGNSVIEENE) from the C-terminus of the rat EAAT1 sequence (Rothstein et
al., 1995).
Cell fractionation. Media was aspirated from human fibroblasts grown to 80-90%
confluence in 150-mm dishes. Cells were washed three times with 10 ml ice cold PBS or
SEB (85.6 g/L sucrose. 0.76 g/L EGTA. 2.38 g/L Hepes, pH 7.5). then scraped and
collected in a plastic 50 ml centrifuge tube. A total of 15 ml of SEB with 0.5 mM
phenylmethyl sulfonyl fluoride (PMSF) and 1 pl/ml protease inhibitors (1 pg/ml antipain
and leupeptin. and 100 KIU/ml aprotinin) per 150-mm dish was used to scrape the cells.
Using a refrigerated table-top centrifuge, cells were spun at 300 x g for 5-10 minutes,
supernatant discarded, and cell pellets resuspended in 15 ml of SEB + PMSF and protease
inhibitors. Cell suspension was poured into an ice-cold nitrogen bomb (Parr Instrument
Company, Moline. IL) and allowed to equilibrate at 200 psi for 10 min before lysis by
rapid release. Resultant homogenate suspension was collected in a centrifuge tube and
checked for unbroken cells under a light microscope. Homogenate was centrifuged at
300 x g for 10 min as described above, and the pellet was saved as the nuclear fraction.
The 300 x g supernatant was spun in an ultracentrifuge at 15.000 x g (14.500 rpm in a
Beckman 60TI rotor) for 30 min and the pellet saved as a crude plasma membrane-
enriched fraction. The supernatant from the 15,000 x g spin was centrifuged in the
ultracentrifuge at 100.000 x g (37.500 rpm in a Beckman 60TI rotor) for 60 min and the
pellet was saved as the total intracellular membrane fraction.
SDS-PAGE and immunoblot analysis. Samples from the cell fractionation
procedure were initially dissolved in 0.2 N NaOH/O.2% SDS, and then further diluted to


85
microtubules appeared to be intact in both the normal (Figure 4-7 E) and LPI cells
(Figure 4-7 F) according to immunofluorescent labeling using a 1:100 dilution of mouse
anti-P-tubulin antibody.
Antibodies against resident proteins of the endoplasmic reticulum (ER) and Golgi
were used to compare the organelles of the biosynthesis pathway in normal and LPI cells.
A short, four amino acid (Lys-Asp-Glu-Leu) peptide, called KDEL, appears in the
sequences of ER resident proteins (such as BiP) and is responsible for selectively
retrieving the proteins after they leave the ER in transport vesicles (reviewed by Pelham,
1991). A membrane-bound receptor in the cis-Golgi recognizes the KDEL retention
signal, and returns the KDEL-containing proteins to the ER. The ER, in normal and LPI
fibroblasts, was detected using a 1:100 dilution of mouse anti-KDEL antibody (Figure 4-
8 A and B), a 1:50 dilution of rabbit anti-BiP antibody (data not shown), as well as, a
1:25 dilution of rabbit anti-protein disulfide isomerase (PDI) antibody (data not shown).
The KDEL antibody provided the best labeling, but each of the antibodies stained the ER
with no significant difference between the two cell lines. The ER in the LPI cells
appeared to be larger and more spread out than the ER of the normal fibroblasts;
however, this was probably due to the fact that the LPI cells tend to be larger in general.
Compartments of the Golgi were stained using 1:50 dilutions of either rabbit anti-
sialyltransferase antibody (Figure 4-8 C and D), which labels the Golgi and Trans-Golgi
Network (TGN), or rabbit anti-mannose-6-phosphate receptor antibody (Figure 4-8 E and
F), which is specific for TGN and late endosomes. There was no recognizable difference


15
Cell Biological Techniques for Studying Protein Trafficking
Indirect immunofluorescence has provided a way of studying the subcellular
localization and trafficking of proteins using a combination of antibodies and trafficking
inhibitors. Indirect immunofluorescence is a sensitive method for detecting a protein of
interest because many molecules of the secondary antibody recognize each molecule of
primary antibody, which is raised against a specific peptide or protein. This results in an
amplification of the signal because the secondary antibody is covalently attached to a
fluorochrome that fluoresces when exposed to a specific wavelength of light. However,
problems may arise if an antibody cannot be raised against a desired protein, or if an
antibody produces a high level of background by cross-reacting with other cellular
proteins or artifacts. To avoid some problems commonly associated with antibodies, and
to investigate living cells, a new technique that utilizes the autofluorescence of the green
fluorescent protein (GFP) has replaced, or is being used in conjunction with, antibody
labeling techniques.
GFP. a 27 kDa protein native to the bioluminescent jellyfish, Aequorea victoria,
produces a bright green color when stimulated by blue or UV light (reviewed by Steams,
1995). GFP expression is species-independent and can be introduced into prokaryotic
and eukaryotic cells without the requirement of specific cofactors, substrates, or
additional gene products. The GFP protein is small and globular, and in most cases does
not interfere with the synthesis, trafficking, or activity of the fusion protein product.
Whereas antibody labeling often involves the use of a fixative, GFP constructs can be


39
conjugated to fluorescein isothiocyanate (FITC). For double-labeling experiments,
monoclonal antibodies against the organelle-specific proteins were detected using a 1:200
dilution of goat anti-mouse IgG linked to Texas Red. The results of all
immunofluorescence experiments in this chapter were analyzed using deconvolution
microscopy (described
in the Methods Chapter).
Table 3-1
Antibodies against Glutamate Transporters
Name
Host
Source
Dilutions
jp* IB**
EAAT3
rabbit
Dr. Michael Kilberg.
Univ. of Florida
1:200
--
EAAT1-R
rabbit
Dr. Jeffrey Rothstein.
Johns Hopkins
1:50
1:200-
1:1000
EAAT1-S
rabbit
Dr. Wilhelm Stoffel,
Univ. of Cologne, Germany
1:50
1:50
EAAT1-C
guinea pig
Chemicon International.
Temecula, CA
1:1000
1:100
EAAT1-D
rabbit
a-Diagnostics,
San Antonio, TX
1:50

KDEL
mouse
Dr. David Vaux,
EMBL
1:100

Transferrin receptor
mouse
Zymed,
San Francisco, CA
1:5

414
mouse
Dr. John Aris,
Univ. of Florida
1:10

D77
mouse
Dr. John Aris,
Univ. of Florida
1:50

*IF = immunofluorescence assay
**IB = immunoblot
Results
Localization of endogenous EAAT3 glutamate transporter in human fibroblasts by
immunofluorescence. Immunofluorescence assays were performed on human fibroblasts
using antibodies generated against the glutamate transporters, EAAT1 and EAAT3.


134
Co-localization of CAT1 and eNOS within PAEC. Several laboratories have
documented the co-localization of caveolin and eNOS by immunohistochemistry, and
shown that eNOS is co-immunoprecipitated with anti-caveolin antibodies (Garcia-
Cardena, 1996; Feron et al., 1996). When these co-localization experiments were
reproduced by myself using PAEC and human fibroblasts, caveolin and eNOS showed
significant overlap in both cell types (data not shown). These results, in addition to the
presence of CAT1 in the caveolae, raised the possibility that CAT1 and eNOS may be co
localized in a caveolar complex to facilitate NO synthesis. When paraformaldehyde-
fixed PAEC were subjected to immunohistochemistry to test for co-localization of CAT1
and eNOS, a significant portion of the detectable membrane-associated eNOS was
clustered within specific regions on the cell surface (Figure 5-3 C). The CAT1 antibody
was detected with a goat anti-rabbit IgG conjugated to FITC, and eNOS was visualized
with a goat anti-mouse IgG linked to Texas Red. Although clearly separate pools also
exist, many of the eNOS-labeled membrane clusters overlapped with the CAT1 staining.
A large quantity of eNOS and CAT1 were detected in a perinuclear region
resembling the Golgi. To confirm that the perinuclear staining was Golgi, both eNOS
and CAT1 were used in double-labeling experiments with Golgi-specific antibodies.
When MeOH-fixed PAEC were labeled with a 1:50 dilution of CAT1 and a 1:100
dilution of the cis-Golgi marker, 10e6 (see Table 5-1), the CAT1 antibody demonstrated
strong co-localization with 10e6 in the Golgi region (Figure 5-4 A). Likewise, eNOS
(1:10 dilution) and the Golgi-marker, sialyltransferase (1:50 dilution), showed significant
overlap in the perinuclear region of MeOH-fixed PAEC (Figure 5-4 B). These data are
consistent with the report by Sessa et al. that a significant amount of eNOS resides in the


143
specifically directing eNOS to the caveolae, however, other mechanisms may be involved
in stimulating the trafficking when there is a need for NO production. Unpublished
observations in the laboratory of Dr. Edward Block, University of Florida, suggest that
reducing the amount of extracellular L-arginine reduces the amount to NO produced by
the PAEC. Also, preliminary results have shown that blocking the uptake of L-arginine,
using lysine as a competitive inhibitor, blocks the production of NO. To investigate
whether or not extracellular arginine concentrations detectably influence the amount of
eNOS on the plasma membrane, PAEC were incubated in complete MEM + 10% FBS
(Figure 5-11 A), complete MEM minus L-arginine + 10% FBS (Figure 5-11 B), or
complete MEM plus 0.5 mM L-arginine + 10% FBS (Figure 5-11 C). One set of cells
was incubated in the different media solutions for 1 h (under normal culture conditions,
see Methods Chapter), and a second set of cells was incubated overnight. Cells were then
fixed with -20C MeOH and stained with anti-eNOS antibody. Regardless of the
concentration of L-arginine in the media, the eNOS antibody staining was identical, with
a small amount of eNOS detected on the plasma membrane and abundant perinuclear
staining. These results suggest that extracellular L-arginine is not a factor in the
recruitment of eNOS from the Golgi to the plasma membrane in cultured PAEC.
Discussion
The clinical importance of the vascular L-arginine/nitric oxide synthase cycle is
validated when considering the number of pathophysiological conditions that arise from a
disruption of the system. Alterations in the production or availability of NO have been
implicated in contributing to complications arising in such diverse disorders as


87
large "vacuoles" that were previously observed by phase contrast and electron
microscopy. Acridine orange (AO) is a lysosomotropic weak base that accumulates in
acidic compartments of living cells (Robbins et al 1963). When excited with blue light,
it emits a red fluorescence that can be visualized using a Texas Red filter. To determine
whether or not the lpgl20-containing compartment was acidic in nature, as would be
expected for lysosomes, normal and LPI fibroblasts were loaded with 5 pg/ml acridine
orange for 15 min before fixing with either 4% paraformaldehyde or -20C MeOH and
viewing with a Nikon axiophot inverted epifluorescent microscope (Figure 4-12 A and
B). Deconvolution microscopy could not be performed because the acridine orange
diffused out of the acidic compartments too rapidly. For the same reason, staining with
the lpgl20 antibody after loading with acridine orange was unsuccessful. However, it did
appear as though the AO-containing compartment in the LPI cells was larger and more
abundant (Figure 4-12 B), even though it was not confirmed that these structures
represented the lpgl20-containing lysosomes. The same AO staining patterns were
obtained when live normal or LPI cells were loaded with AO and immediately viewed
with the epifluorescence microscope. When exposed to an extremely high dose of AO
(ranging from 50-500 pg/ml), the normal fibroblasts died immediately, as judged by the
way they curled and lifted off the tray, whereas the LPI cells were able to tolerate the
drug (data not shown).
The lysosomes of normal and LPI fibroblasts were also compared following
treatment with chloroquine, a lysosomotropic drug that neutralizes lysosomes and other
acidic compartments. Both cell lines were incubated with 50 pM of chloroquine in MEM


93
fibroblasts (Davies et al., 1997). Rab7 resides in late endosomes and is believed to
participate in the trafficking of proteins from early to late endosomes. Rab9 has been
implicated in the recycling of the mannose-6-phosphate receptor from late endosomes to
the TGN. Because of their involvement in the recycling and degradative pathways, as
well as, the morphological changes in lysosomes upon their down-regulation, Rab7 and
Rab9 could be candidates for association with the CHS/Beige protein. CHS patients
suffer from severe immunologic defects, abnormal platelet function, and partial ocular
and cutaneous albinism (Perou et al., 1997). Although the clinical symptoms of LPI are
different from those of CHS, the perinuclear accumulation of abnormally large vesicles in
the LPI fibroblasts is remarkably similar to the abnormal lysosomal vacuoles of CHS
fibroblasts.
There are several lysosomal storage disorders that result from the abnormal
accumulation of small molecules in the lysosomes (reviewed by Chou et al., 1992). The
sialic acid storage diseases represent a collection of inherited disorders characterized by
the elevated excretion of sialic acid in the urine, as well as, the accumulation of the acidic
monosaccharides in lysosomes (Mancini et al., 1991). Mancini and coworkers identified
and characterized a proton-driven sialic acid carrier with broad specificity to other acidic
monosaccharides in the lysosomal membranes of rat liver (Mancini et al., 1991).
Therefore, a defect in this carrier may represent the molecular basis underlying one or
more of the sialic acid storage diseases.
A number of amino acid transport systems have also been characterized in the
lysosomal membranes of human fibroblasts and rat liver (Chou et al., 1992). System c
represents a carrier-mediated cationic amino acid lysosomal transporter that shares many


61


161


84
described in the literature for a variety of cell lines. Initial experiments were performed
in order to confirm the reliability of the antibodies and document the staining patterns in
the normal human fibroblasts. For all of the immunoassays, secondary antibodies used
were either goat anti-rabbit IgG or goat anti-mouse IgG conjugated to either fluorescein
isothiocyanate (FITC, Sigma Chemical Co., St. Louis, MO) or Texas Red (TR, Cappel
Laboratories, Durham, NC).
The nuclear membranes of normal (Figure 4-6 A) and LPI (Figure 4-6 C) cells
were examined using a 1:10 dilution of mouse anti-human 414 antibody, which was
generated against an epitope common to several of the nuclear pore complex proteins
(Davis and Blobel, 1986). A 1:50 dilution of mouse anti-yeast D77 antibody (Aris and
Blobel, 1988) was used to visualize the nucleoli of both normal (Figure 4-6 B) and LPI
(Figure 4-6 D) fibroblasts. The labeling of the nuclear structures was consistent with
images from the literature, and there was no discernible difference between the staining
patterns of the normal and LPI fibroblasts. A 1:100 dilution of mouse anti-human-p-
integrin antibody was used to label the plasma membrane (Figure 4-7 A and B), and a
1:200 dilution of rabbit anti-caveolin-1 antibody was used to specifically detect the
caveolar domains of the plasma membrane (Figure 4-7 C and D). In both normal and LPI
cells, the caveolin antibody labeled specific regions of the plasma membrane, as
expected, and the P-integrin antibody showed strong staining around the periphery of
both cell types. Although the LPI cells may be defective in cell-to-cell contact, indicated
by the way they grow in clusters rather than spreading out to confluence, there was no
noticeable difference in the P-integrin staining pattern of the normal and LPI cells. The


142
demonstrated earlier, both the caveolin and eNOS antibodies co-localize with the
expressed wild-type GFP(C3)-CAT1 (Figure 5-6). However, the GFP(C3)-CAT1 is
expressed at a much higher level than the endogenous, therefore, there is a significant
amount of the fusion protein on the plasma membrane that does not co-localize with
either caveolin or eNOS. CATMUT1 was expressed normally, trafficked to the plasma
membrane, and co-localized with both caveolin (Figure 5-9 A) and eNOS (Figure 5-9 B)
to the same degree as the wild-type fusion protein. CATMUT2 and CATMUT3 also
expressed well and were detected on the plasma membrane, however, CATMUT3
showed very little co-localization with either caveolin (Figure 5-10 A) or eNOS (Figure
5-10 B). It is hard to determine whether or not there is a true difference between the
wild-type and CATMUT3, though, because the amount of co-localization following the
transfections was low in the wild-type. Also, the transfected cells had a reduced amount
of eNOS on the plasma membrane, so the majority of the co-localization between
GFP(C3)-CAT1 and eNOS was in the perinuclear region. Antibodies against caveolin
and eNOS inconsistently co-localized with the expressed CATMUT2 protein, in that,
some cells would detect overlapping of the two signals and others would not (data not
shown). Whereas this may indicate an intermediate result, no conclusive statements
following three individual experiments can be made concerning the CATMUT2
construct.
eNOS distribution following exposure to extracellular L-arginine. As illustrated
in Figure 5-11, a significant amount of eNOS resides in the Golgi of PAEC. However,
for enzyme activation and maximal NO production, eNOS must be translocated to the
caveolar regions (Liu et al., 1996; Feron et al., 1996). Acylation is likely involved in


130
Table 5-1
Antibodies for Immunofluorescence Studies
Name
Host
Source
Dilutions
IF* IB**
CAT1
rabbit
Dr. Michael Kilberg,
1:50 -
Univ. of Florida
caveolin
mouse
Transduction Laboratories
1:50 -
eNOS
mouse
Lexington, KY
Transduction Laboratories
1:10 1:200
Lexington, KY
(3-tubulin
mouse
Sigma,
1:100 -
St. Louis, MO
sialyltransferase
rabbit
Dr. William Dunn,
1:50 -
Univ. of Florida
10e6
mouse
Dr. William Brown,
1:100 -
Cornell Univ.
*IF = immunofluorescence
**IB = immunoblotting
Immunodepletion of CAT1 transport activity. By collaborators in the laboratory
of Dr. Edward Block, plasma membrane vesicles were prepared by sucrose gradient
centrifugation as described by Teitel (Teitel, 1986) and modified by Bhat and Block
(Bhat and Block, 1990; Bhat and Block, 1992). Plasma membrane proteins were
solubilized by the method described by Fafournoux et al. (Fafoumoux et al., 1989). The
solubilized proteins in the supernatant were precipitated by incubation with 20%
polyethylene glycol (PEG-8000) at 4C for 20 min. Immunodepletion of CAT1
transporter was performed using the protocol of Tamarappoo et al. (Tamarappoo et al.,
1992). Briefly, a 1-ml aliquot of goat anti-mouse IgG covalently linked to agarose beads
(Sigma Chemical Co., St. Louis, MO) was incubated for 1 h with 20 pg of a mouse anti
human eNOS antibody on ice and spun for 5 min at 1,500 x g, after which the supernatant
was discarded. The agarose beads were then washed once with STAB buffer (20%
glycerol, 2 mM EDTA, 2 mM DTT, 0.2% sodium cholate, 0.25% asolectin, and 10 mM


72
unpublished results). The lack of a large difference in initial uptake rates suggests that
influx at the plasma membrane is unaffected by the disease. To eliminate the possible
confounding effects of trans-stimulation when measuring transport in whole cells, a crude
mixture of cellular membrane vesicles was isolated from cultured fibroblasts and used to
assay lysine uptake. Qualitatively, the data obtained from vesicle transport experiments
were similar to the results of whole cell transport. Once again, initial measurements
showed little difference in uptake by normal and LPI-derived vesicles, whereas LPI-
derived vesicles eventually accumulated four to five times more 'H-lysine than the
control vesicles (Handlogten and Kilberg, unpublished results). These results suggested a
decreased efflux of amino acid from either the plasma membrane or intracellular vesicles
of the LPI cells.
The observation that the LPI-derived vesicles also accumulate select neutral
amino acids (Handlogten and Kilberg, unpublished data) is consistent with the detection
of increased neutral amino acids in the urine of LPI patients. This prompted our
laboratory to investigate amino acid transport by other known transport systems in LPI
fibroblasts and plasma membrane vesicles. System b0 + is a NaMndependent system that
mediates the bidirectional transport of both cationic and neutral amino acids (Van
Winkle, 1988; Van Winkle et ah, 1988). However, this system is probably not
responsible for the elevated Na-independent uptake of neutral amino acids observed
during in vitro studies, because the transport of leucine was poorly inhibited by lysine in
the plasma membrane vesicles derived from LPI fibroblasts (Handlogten and Kilberg,
unpublished results). The NBAT protein, which exhibits the properties of System b0 +,


Figure 3-9. EAAT1 immunofluorescent staining of PAEC transfected with EAAT1-
GFP(N3). PAEC were transfected for 3 h with the EAAT1-GFP(N3) fusion protein
according to the lipofectamine protocol described in the Methods Chapter. Following 24-
48 h of expression, cells were fixed with -20C MeOH and stained with antibodies
against EAAT1-R (A) and EAAT1-S (B). Both EAAT1 antibodies were detected with a
goat anti-rabbit IgG conjugated to Texas Red and visualized by deconvolution
microscopy. Images were processed from three independent experiments and the staining
was determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.


138
Immunodepletion of CAT1 activity with an eNOS-specific antibody. Co
localization of the CAT1 arginine transporter and eNOS within caveolae is consistent
with the proposal that these membrane micro-domains are a site for concentrating
proteins involved in signaling (Parton, 1996; Simionescu and Simionescu, 1987;
Schnitzer et al., 1994; Lisanti et al., 1994). However, the present observations also raise
the intriguing possibility that the CAT1 arginine transporter and eNOS are physically
associated. Such a complex might provide a mechanism for directed delivery or even
channeling of newly acquired extracellular arginine to eNOS for NO synthesis. This
channeling would explain the "arginine paradox" whereby eNOS apparently
preferentially accepts extracellular arginine in the presence of saturating intracellular
substrate levels (Kurz et al., 1997). The saturable Na+-independent arginine transport
activity in PAEC vesicles has been characterized previously and is mediated primarily by
the CAT1 transporter given that neither b0,+ nor y+L are detectable in these assays
(Zharikov and Block, 1997). It has also been determined that PAEC contain CAT1, but
little or no CAT2 or CAT2a mRNA (unpublished results). As a test for a direct complex
between CAT1 and eNOS, PAEC plasma membrane arginine transport activity was
detergent solubilized and subjected to immunoprecipitation with anti-eNOS antibody, as
described in the Methods Chapter (Fafournoux et al., 1989). Using the proteins in the
immunodepleted supernatant to reconstitute proteoliposomes (Tamarappoo et al., 1992)
provided an assay to check for anti-eNOS dependent immunodepletion of arginine
transport. Immunodepletion with non-immune mouse IgG caused no loss of
reconstitutable arginine transport, whereas the anti-eNOS monoclonal antibody caused
immunoprecipitation of 73% of the NaMndependent arginine transport (Table 5-2).


132
Reconstitution and assay of amino acid transport. Reconstitution of soluble
proteins into proteoliposomes was performed following the protocol of Fafoumoux et al.
(Fafoumoux et al., 1989) and transport assays were performed as described previously
(Zharikov and Block, 1997). Briefly, plasma membrane vesicles or proteoliposomes (20
pg/30 pi) in SMB buffer (250 mM sucrose, 1 mM MgCl2, 10 mM Flepes, pH 7.5 ) were
added to 270 pi of external solution containing 140 mM NaSCN, 1 mM MgS04, 10 mM
HEPES-Tris (pH 7.4), and 50 pM [3H]-arginine or 50 pM [3H]-glutamine. After
incubation for 3 min at 37C, reactions were terminated by the addition of 5 ml of ice-
cold 140 mM NaCl (stop solution) followed by filtration through glass-fiber Whatman
GF/C filters presoaked in 0.3% polyethylenimine to decrease the nonspecific absorption
of [3H]-L-arginine or glutamine. The filters were washed 4 times with 5 ml of stop
solution, dried, and trapped radioactivity determined using liquid scintillation
spectrometry. Zero-time blank values (membrane vesicles or proteoliposomes added
after stop solution) were subtracted from all experimental values. The data represent
triplicate assays performed on at least two independent preparations, and are presented as
pmol.mgTprotein.3min'1. Statistical significance was established by Students T test.
Results
Characterization of CAT1 staining of PAEC plasma membranes. Using a CAT1
antibody, generated in our laboratory, in conjunction with immunostaining analysis using
an epifluorescence microscope, it was previously shown that the arginine transporter is
concentrated in specific regions of the plasma membrane in a number of cell types,
including PAEC (Woodard et al., 1994). These results were confirmed and extended


Figure 5-1. Surface labeling of PAEC with the CAT1 transporter antibody. PAEC were
fixed with 4% PFA and subjected to immunohistochemistry using the procedures
described in the Methods Chapter. Panels A and B demonstrate the clustering of the
CAT1 transporter on PAEC using a polyclonal anti-CAT 1 antibody detected with a goat
anti-rabbit IgG conjugated to FITC. Panel A shows staining of PAEC before
deconvolution to reveal the outline of the entire cell body. Panel B shows the same cells
after deconvolution. Panel C demonstrates cells that were incubated with anti-CAT 1
antibody following preadsorption for 12 h at 4C with 50 pg/ml of the corresponding
peptide antigen. Staining from three independent experiments was analyzed by
deconvolution microscopy and shown to be reproducible. The data shown represent
analysis of 0.2 pm sections through the cells. One cell in panel C is outlined to
distinguish the cell periphery.


176
Davies, J. P., Cotter, P. D. and Ioannou, Y. A. (1997) Genomics 41, 131-134
Davis, K. E., Straff, D. J., Weinstein, E. A., Bannerman, P. G., Corrale, D. M.,
Rothstein, J. D. and Robinson, M. B. (1998) J. Neurosci. 18, 2475-2485
Davis, L. I. and Blobel, G. (1986) Cell 45, 699-709
De Silva, D. M Askwith, C. C. and Kaplan, J. (1996) Physiol Rev. 76, 31-44
de Weerd, W. F. and Leeb-Lundberg, L. M. (1997) J. Biol. Chem. 272, 17858-17866
Deves, R., Chavez, P. and Boyd, C. A., R. (1992) J. Physiol., Lond. 454, 491-501
Drumm, M. L., Wilkinson, D. J., Smit, L. S., Worrell, R. T., Strong, T. V., Frizzell, R.
A., Dawson, D. C. and Collins, F. S. (1991) Science 254, 1797-1799
Eddahibi, S., Adnot, S., Carville, C., Blouquit, Y., and Raffestin, B. (1992) Am. J.
Physiol. 263, L194-L200
Fafournoux, P., Dudenhausen, E. E., and Kilberg, M. S. (1989) J. Biol. Chem. 264,
4805-4811
Fairman, W. A., Vandenberg, R. J., Arriza, J. L., Kavanaugh, M. P. and Amara, S. G.
(1995) Nature 375, 599-603
Feron, O., Belhassen, L., Kobzik, L., Smith, T. W., Kelly, R. A., and Michel, T. (1996)
J. Biol. Chem. 271,22810-22814
Feron, 0., Smith, T. W Michel, T. and Kelly, R. A. (1997) J. Biol. Chem. 272, 17744-
17748
Forstermann, U., Closs, E. I., Pollock, J. S., Nakane, M., Schwarz, P., Gath, I., and
Kleinert, H. (1994) Hypertension 23, 1121-1131
Garcia-Cardena, G., Fan, R Stem, D. F., Liu, J., and Sessa, W. C. (1996) J. Biol. Chem.
271, 27237-27240
Garcia-Cardena, G., Oh, P., Liu, J., Schnitzer, J. E., and Sessa W. C. (1996) Proc. Natl.
Acad. Sci. USA 93, 6448-6553
Ginsberg, S. D., Martin, L. J. and Rothstein, J. D. (1995) J. ofNeurochem. 65, 2800-
2803


146
evidence, but those experiments were negative. It is possible that the eNOS antibody is
better suited for the immunoprecipitation protocol than the CAT1 antibody. The anti-
CAT 1 does not immunoblot effectively either. In the future, problems involving the
CAT1 antibody may be overcome by overexpressing the GFP(C3)-CAT1 in PAEC, and
then immunoprecipitating eNOS using an anti-GFP antibody. The CAT1 transporter may
be in such low abundance that the amount precipitated with eNOS is below the level of
detection of the immunoblot (especially, if the interaction between CAT1 and eNOS is
transient). Michel et al. proposed a model of reciprocal regulation in which eNOS binds
caveolin and the calcium/calmodulin complex via the same or overlapping sequences
(Michel et al., 1997). eNOS is inactive when it is bound to caveolin. Intracellular
concentrations of calcium are increased following agonist stimulation, allowing calcium
to bind, and activate, calmodulin. The calcium/calmodulin replaces caveolin and eNOS
is activated. It is possible that activated eNOS is allowed to diffuse freely within the
caveolar domain, interact transiently with CAT1 in order to produce NO, then return to
caveolin and the inactive state. Alternative approaches may address these questions. The
yeast two-hybrid system has proven to be a sensitive assay for detecting interactions
between proteins that fail to co-immunoprecipitate and could be applied to CAT1 and
eNOS. For instance, Ras and Raf did not co-precipitate together; however, an association
was detected using the yeast two-hybrid system (Van Aelst et al., 1993). The Kilberg
laboratory has initiated a collaboration with Dr. Bettie Sue Masters, Southwestern
Medical School, to test for binding between CAT1 peptide sequences and purified eNOS.
Association of the CAT1 arginine transporter and eNOS in PAEC provides a
mechanism for the directed delivery of L-arginine substrate to eNOS. NO is a highly


Figure 3-1. Extracellular staining of human fibroblasts with EAAT3 antibody. Human
fibroblasts were fixed with 4% PFA and subjected to immunohistochemistry using the
procedures described in the Methods Chapter. Panel A demonstrates the clustering of
EAAT3 transporters on the surface of human fibroblasts using a polyclonal anti-EAAT3
antibody detected with a goat anti-rabbit IgG conjugated to FITC. Panel B shows cells
that were incubated with anti-EAAT3 antibody following preadsorption for 12 h at 4C
with 50 pg/ml of the corresponding peptide antigen. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
An outline was drawn around one cell to distinguish the periphery. The data shown
represent analysis of 0.2 pm sections through the cells.


Figure 3-2. Intracellular staining of human fibroblasts with EAAT3 antibody and co
localization with organelle-specific antibodies. Using the methods described in the
Methods Chapter, human fibroblasts were fixed with -20C MeOH and subjected to
immunohistochemistry with antibodies specific for EAAT3 (A), EAAT3 and KDEL (B),
or EAAT3 and transferrin receptor (TfR) (C). Co-localization of the proteins was
assayed by using simultaneously a rabbit polyclonal antibody against EAAT3 detected by
FITC-labeled goat anti-rabbit IgG and mouse monoclonal antibodies against KDEL and
TfR detected by Texas Red labeled goat anti-mouse IgG. Staining from three
independent experiments was analyzed by deconvolution microscopy and shown to be
reproducible. The data shown represent analysis of 0.2 pm sections through the cells.


CHAPTER 6
CONCLUSIONS AND FUTURE DIRECTIONS
Until the late 1980s, the process of amino acid transport into mammalian cells
was attributed to a loosely-defined group of "systems" with overlapping properties.
Substrate specificity, many regulatory properties, and kinetic parameters were well-
characterized, yet the actual proteins responsible for the activities remained unknown.
Within the last decade, more than two dozen amino acid transporters have been
identified, cloned, and categorized into families according to intrinsic structural and
functional properties. Mutagenesis studies have provided information concerning
substrate binding and transport activity. Northern analyses and immunoblots have
provided evidence concerning the differential expression and regulation of members of
the same transporter family. However, very little is known regarding the "life cycle" or
the active trafficking of amino acid transporters. It is believed that amino acid
transporters follow many of the individual steps, from biosynthesis to degradation, that
have been delineated for other plasma membrane proteins. With antibodies and
molecular tags available, it has been possible to study the subcellular localization and
trafficking of specific amino acid transporters. The individual projects that have
contributed to this thesis share a common interest in documenting the cellular
localization, and when possible, understanding the functional consequences of transporter
protein distribution.
169


117
15.O um


eoi


4-6. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the nuclear membrane and nucleolus 107
4-7. Intracellular staining of normal and LPI fibroblasts with antibodies
against plasma membrane and cytoskeletal proteins 109
4-8. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the endoplasmic reticulum and Golgi Complex Ill
4-9. Intracellular staining of normal and LPI fibroblasts with antibodies
against proteins of the endocytic and recycling pathways 113
4-10. Intracellular staining of normal and LPI fibroblasts with an
antibody against a lysosomal enzyme 115
4-11. Intracellular staining of normal and LPI fibroblasts with an
antibody against a lysosomal membrane protein 117
4-12. Visualization of acidic compartments of normal and LPI
fibroblasts with acridine orange 119
4-13. Lysosomal detection in normal and LPI fibroblasts following
chloroquine treatment 121
4-14. Lysosomal staining of normal and LPI cells expressing the
GFP(C3)-CAT1 fusion protein 123
5-1. Surface labeling of PAEC with the CAT1 transporter antibody 149
5-2. Co-localization of CAT1 and caveolin on PAEC 151
5-3. Co-localization of CAT1 and eNOS on PAEC 153
5-4. Detection of CAT1 and eNOS in the Golgi of PAEC 155
5-5. Expression of the GFP(C3)-CAT1 fusion protein in PAEC 157
5-6. Immunofluorescent staining of PAEC transfected with GFP(C3)-
CAT1 fusion protein 159
5-7. Distruption of CATl/eNOS co-localization in PAEC treated with
nocodazole 161
5-8. CAT1 immunoprecipitation of eNOS from solubilized PAEC
plasma membrane vesicles 162
vii


23
Quiagen or Boehringer Mannheim and proved to be the least cytotoxic and provide the
greatest transfection efficiency (about 15-20%). To remove endotoxins from the cDNA
preps. cDNA in solution was mixed with 1% Triton X-l 14. vortexed, and chilled on ice
for 5 min. The sample was heated for 5 min at 37C, and centrifuged at 14,000 x g for 5
min before recovering the aqueous solution to be used in the transfection protocols.
Preparation of transporter cDNA-Green Fluorescent Protein constructs. The
pEGFP vectors from Clontech (Palo Alto, CA.) encode the Green Fluorescent Protein
(GFP) variants that fluoresce 35 times more intensely than wild-type and have been
codon-optimized for maximal translation efficiency in mammalian cells (Cormack et al.,
1996). These vectors are available in all three reading frames and contain 20 unique
restriction sites in the multiple cloning region to facilitate subcloning. The pEGFP(N3)
Protein Fusion Vector from Clontech was used to fuse the EAAT1 cDNA to the N-
terminus of EGFP(N3). The EAAT1 cDNA in pCDNA3 was obtained from Dr. Jeffrey
Rothstein's laboratory. PCR primers were designed to the 5 terminus beginning at the
ATG translation start site (ATGACTAAAAGCAATGGAGAAGAGC) and the 3
terminus ending at nucleotide 1680 (CATCTTGGTTTCACTGTCGATGG). Using 5 ng
of EAAT1 cDNA and 100 pmol of each primer, the entire coding region minus the stop
codon was amplified with Taq polymerase according to the manufacturers protocol
(InVitrogen, Carlsbad, CA). Amplification proceeded for 30 cycles of the following
conditions: denaturation at 94C for 1 min. annealing at 45C for 1 min, and extension at
72C for 1 min, with a final extension of 10 min. After the PCR product was obtained,
the 1680 base pair band was gel purified according to the protocol of Quiagen (Valencia,
CA) and cloned into the TA cloning vector, pCR 2.1 according to the manufacturer's


32
transporter (Rothstein et al.. 1995). It has recently been proposed that the down-
regulation of EAAT2 results from a defect in mRNA processing. Lin et al. has shown
that due to defective mRNA splicing events such as intron-retention or exon-skipping,
multiple abnormal EAAT2 mRNA species are produced in the affected areas of the brain
in ALS patients (Lin et al.. 1998). In vitro expression studies suggested that the protein
products of the aberrant mRNAs had decreased transport activity because they were
degraded rapidly, or perhaps, had a dominant-negative effect on the normal EAAT2
protein. The abnormal mRNA species were not present in regions of the brain that were
unaffected in ALS patients, or in the non-neurologic disease controls.
Decreased glutamate transporter activity in the frontal, parietal, and temporal
cortex has been implicated in the neurodegeneration that occurs in Alzheimer disease
(Scott et al.. 1995; Cowbum et al.. 1988). Like ALS. mRNA levels of EAAT1. EAAT2.
and EAAT3 were normal in the frontal cortex, however, immunoblot analysis detected
about 30% less EAAT2 protein (Li et al.. 1997). On the other hand, schizophrenia and
other psychoses are thought to result partially from glutamatergic hypofunction. a
condition that occurs following excessive glutamate uptake (Carlsson and Carlsson.
1990). Therefore, the mechanism by which glutamate is cleared from the synaptic cleft
must be tightly regulated in order to prevent neuronal damage or malfunction.
The roles that the individual transporters play in normal synaptic clearance and
neurotoxicity are unclear because subtype-specific inhibitors are not available. This has
led a number of laboratories to study the functions of these specific transporters using
antisense oligonucleotides or knockout mice. Results from antisense studies indicated
that a loss of EAAT1 and EAAT2 resulted in elevated extracellular glutamate


21
different fluorescent properties, these two secondary antibodies can be used in double
labeling experiments to show co-localization of two different proteins. However, the
proteins of interest must be detected using primary antibodies generated in two different
species. For example, double-labeling can be performed by incubating cells with a
primary antibody raised in mouse, and detected with a secondary goat anti-mouse Texas
Red antibody, and with another primary antibody raised in rabbit, detected using a goat
anti-rabbit FITC antibody.
Fluorescence light microscopy. Slides were initially viewed using a Nikon
Axiophot epifluorescence inverted microscope. A Leitz Planapo 63x, NA/1.4 oil
immersion lens and a modified Zeiss Axiomat inverted light microscope was used for
collecting three-dimensional light microscopy data sets (Agard, 1984; Agard and Sedat,
1983). The focal position, UV excitation shutter, and digital camera shutter of the
microscope were under computer control. The images generated were digitized directly
from the microscope image plane using a 14 bit, liquid nitrogen-cooled charge-coupled
device (CCD) digital camera (described in detail in Hiraoka et al., 1987; Agard et al.,
1989; Paddy et al., 1990). Three-dimensional data sets were collected as a series of
images separated by 0.5 mm along the horizontal optical sectioning axis (this value varies
with the depth of the cell). For double-labeling experiments consisting of different
fluorescence wavelengths, a complete 3-D data set at the first wavelength was collected,
the focus was returned to the starting focal position and the barrier filter was changed
using the computer control, then the second data set was collected (Hiraoka et al., 1991).
After data collection, each 3-D data set was corrected for stage and/or sample drift,
fluorescence photo-bleaching through the data set, and lamp intensity and/or shutter open


171
membrane protein, lpgl20. In the future it will be important to isolate the lysosomes
from normal and LPI fibroblasts for further characterization. The sizes of the lysosomal
vesicles isolated from normal and LPI cells can be measured and compared by flow
cytometry. Additionally, measuring the pH will determine if the abnormality is linked to
a change in the acidity of the lysosomal compartment in LPI cells. There is no evidence
to suggest that the CAT1 plasma membrane transporter is contained in the abnormal LPI
vesicle population, however, we hypothesize that the vesicles may accumulate elevated
levels of cationic amino acids via the lysosomal cationic amino acid transport system
(system c). Performing transport studies that compare the influx and efflux of various
amino acids in the normal and LPI lysosomal preparations would be necessary to
investigate this hypothesis. There are several possibilities that may result in the abnormal
size of the LPI vacuoles. As mentioned above, accumulation of amino acids, other small
molecules, or fluids may result in the swelling of the vacuoles. Alternatively, the
contents of the vesicles may be normal, and the enlarged size due to accelerated
membrane fusion or reduced membrane fission. It will be important to determine which
of the above explanations, if any, applies to the abnormal LPI vesicle morphology.
Chapter 5 explores beyond the distribution of amino acid transporters, and
proposes functional consequences for the specific localization of the CAT1 amino acid
transporter in endothelial cells. This project was an extension of previous experiments in
our laboratory by Dr. Michelle Woodard, who first observed the CAT1 transporter
"patching" pattern in several cell types. Co-localization experiments strongly suggested
that CAT1 was localized to plasma membrane and/or Golgi caveolae, where it may be
directly associated with eNOS. If a complex truly exists, it would provide a mechanistic


79
Name
Host
Source
Dilution
M6P receptor
rabbit
Dr. Peter Nissley,
NIH
1:50
Rab 5
rabbit
Santa Cruz Biotechnology
Santa Cruz, CA
1:100
transferrin receptor
mouse
Zymed.
San Francisco, CA
1:5
lpgl20
Rabbit
Dr. William Dunn,
Univ. of Florida
1:100
cathepsin D
Rabbit
Biodesign International,
Kennebunk, ME
1:300
Distribution of endogenous CAT1 in normal and LPI fibroblasts. As mentioned
above, the LPI disease was originally characterized by elevated concentrations of arginine
and lysine in the urine. Later, it was determined that the transport defect is expressed in
the kidney tubules, intestines, cultured fibroblasts, and probably hepatocytes (Simell,
1989). It has been shown that CAT2 and CAT2a are not expressed in human fibroblasts,
so our laboratory began to investigate the localization of the CAT1 transporter in normal
and diseased cells. Normal and LPI fibroblasts were fixed with 4% PFA and labeled with
a 1:25 dilution of the CAT1 transporter antibody (according to the protocol in the
Methods chapter). Antibody staining was detected with a 1:200 dilution of goat anti
rabbit IgG linked to FITC. Both normal and LPI fibroblasts demonstrated an
extracellular periodic labeling that resembled intensely stained patches on the plasma
membrane (data not shown). Staining of the fluorescent patches was completely blocked
when the murine CAT1 antibody was pre-incubated with 50 pg/ml of corresponding
peptide for 12 hours before labeling. This same pattern was observed in several different
LPI fibroblast cell lines, as well as, in porcine pulmonary artery endothelial cells (see
Chapter 5). Incubation of the cells with the microtubule inhibitor, nocodazole, caused the


76
1990; Drumm et al.. 1991). Whereas both the wild-type and mutant CFTR proteins are
associated with calnexin in the ER, only the wild-type protein exits the ER and is
correctly trafficked to the plasma membrane (Pind et al., 1994).
Although no mutations have been found in the CAT1 transporter gene in LPI
patients, or any other protein at this point, the LPI defect results in pleiotropic effects on
amino acid transport (Simell, 1989). Broad scope changes in transport could result from
a block of transporter processing at any stage along the various pathways of membrane
protein synthesis or endocytic recycling. This chapter describes the cellular localization
of the CAT1 arginine transporter and of various organelle-specific proteins along the
biosynthetic, endocytic. and degradative pathways in normal and LPI fibroblasts.
Results
Morphology of normal and LPI fibroblasts by light and electron microscopy.
There are basic differences between the normal and LPI cells that can be discerned at the
light microscope level (Figure 4-1). The LPI cells frequently appear to be larger with
long processes and extensions at the cell periphery. Also, normal fibroblasts grow
significantly faster than their LPI counterparts under the same culture conditions until
they reach confluency. The LPI cells tend to grow in clusters and never reach full
confluence. The most striking difference between the normal and LPI cells is a
population of large vesicles or vacuoles observed throughout the cytoplasm, often
concentrated around the nucleus of the LPI cells. When electron microscopy was used to
increase the resolution of these large intracellular vacuoles, at a magnification of 25K
they appear to contain a fibrous material of unknown origin (Figure 4-2). The unusual


16
viewed in fixed or living cells. Certain GFP variants have been optimized for use in
mammalian cells and are now commercially available. These variants that contain
double-amino-acid substitutions Phe-64 to Leu and Ser-65 to Thr result in a 35-fold
increase in fluorescence over wild type GFP (Cormack et al., 1996). In addition,
expression has been enhanced by the introduction of silent mutations in the coding
sequence that correspond to human codon-usage preferences. The GFP serves as a
genetic tag that can be conveniently added to the protein coding sequence of a cDNA.
In this study, molecular and cell biological techniques, including
immunohistochemistry, GFP expression, and deconvolution microscopy, have been used
to examine specific aspects of the life cycle of amino acid transporters. The individual
projects presented in this thesis share a common goal in documenting the cellular
localization, and when possible, understanding the functional consequences, of
transporter distribution under normal and diseased conditions.


Figure 4-9. Intracellular staining of normal and LPI fibroblasts with antibodies against
proteins of the endocytic and recycling pathways. According to the protocol described in
the Methods Chapter, normal (A and C) and LPI (B and D) fibroblasts were fixed with
-20C MeOH and subjected to immunohistochemistry with the anti-Rab5 antibody
specific for early endosomes (A and B), and the anti-transferrin receptor (TfR) antibody
specific for vesicles involved in recycling (C and D). The anti-TfR primary antibody was
visualized with a goat anti-mouse IgG conjugated to FITC, and the anti-Rab5 antibody
was detected with a goat anti-rabbit IgG linked to FITC. Staining from three independent
experiments was analyzed by deconvolution microscopy and shown to be reproducible.
The data shown represent analysis of 0.2 pm sections through the cells.


44
not shown), however, no nuclear staining was observed in the PAEC (Figure 3-9 A).
Instead, the EAAT1-R antibody stained small vesicles throughout the cytosol of the
PAEC. with slightly more fluorescence concentrated in the perinuclear region. As was
the case for the human fibroblasts, no nuclear staining was detected with the EAAT1-S
antibody in any of the cell lines. In Hela. HepG2, and PAEC, the EAAT1-S antibody
detected only small vesicles of unknown origin in the cytosol. Figure 3-5 (A and B)
shows representative Hela cells stained with the EAAT1-R and EAAT1-S antibodies. _
Identification of EAAT1 glutamate transporter by immunoblot analysis. Total
intracellular membrane (100,000 x g), crude plasma membrane-enriched (15,000 x g),
and nuclear fractions (300 x g) from human fibroblasts, HepG2. and PAEC were isolated,
subjected to SDS-PAGE, and transferred to nitrocellulose membranes for
immunoblotting. The primary antibodies were detected using 1:2.500 to 1:10.000
dilutions of goat anti-rabbit IgG (EAAT1-R and EAAT1-S) or goat anti-guinea pig IgG
(EAAT1-C) conjugated to horseradish peroxidase (described in the Methods Section of
this chapter). When the 300 x g nuclear fraction from human fibroblasts was incubated
with a 1:1000 dilution of EAAT1 -R (Figure 3-6), a strong band was detected at
approximately 70-75 kDa, corresponding to the molecular mass of the monomeric
EAAT1 protein (Tessmann and Kilberg, unpublished data). Less intense bands were
observed at higher molecular masses, including a light band at approximately 180 kDa,
which corresponds to the molecular mass of a putative EAAT1 trimer. EAAT1 was also
detected in the 100,000 x g membrane fraction, however, the protein was primarily
detected at approximately 180 kDa, which is probably the trimeric form. Our laboratory
observes that the more manipulation of the sample, such as the additional centrifugation


Figure 5-7. Distruption of CATl/eNOS co-localization in PAEC treated with
nocodazole. PAEC were incubated with DMSO (A and C) or nocodazole (B and D) for 1
h then fixed with -20C MeOH and stained with an antibody against (3-tubulin (A and B)
or co-stained with antibodies against CAT1 and eNOS (C and D). The (3-tubulin and
eNOS antibodies were detected with goat anti-mouse IgG conjugated to FITC and the
CAT1 antibody was detected with goat anti-rabbit IgG linked to Texas Red. Images were
processed from three independent experiments and the staining was determined to be
reproducible. The data shown represent analysis of 1.0 pm sections through the cells.


68


Figure 5-5. Expression of the GFP(C3)-CAT1 fusion protein in PAEC. PAEC were
transfected for 3 h with the GFP(C3)-CAT1 fusion protein according to the lipofectamine
protocol described in the Methods Chapter. Following 24-48 h of expression, cells were
fixed with -20C MeOH and visualized by deconvolution microscopy. Panel A shows
the expression pattern of the GFP(C3)-CAT1 fusion protein on the plasma membrane and
in intracellular vesicles. PAEC in panel B were stained with the CAT1 antibody
following expression of the GFP(C3)-CAT1 fusion protein. The CAT1 antibody was
visualized with a goat anti-rabbit IgG linked to Texas Red. Images were processed from
three independent transfections and the fluorescence patterns of expressed proteins were
determined to be reproducible. The data shown represent analysis of 0.2 pm sections
through the cells.


131
HEPES, pH 7.4) and mixed with solubilized proteins resuspended in STAB buffer. After
incubation for 1 h on ice, the beads were pelleted as above and discarded, the supernatant
was saved, and the solubilized, non-precipitated proteins were reconstituted into
proteoliposomes.
SDS-PAGE and anti-eNOS immunoblotting. A 125 pi aliquot of sample dilution
buffer (SDB) with (3-mercaptoethanol (BME) was added to the protein A-Sepharose
immunoprecipitated proteins. The samples were heated to 65C for 10 min and then the
sepharose was pelleted by spinning at 13,000 x g for 10 min. Supernatant was removed
from the top and placed in a centrifuge tube and heated again for 10 min at 65C. Each
sample (supernatant from above) was loaded in one lane and subjected to one
dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE)
(Laemmli, 1970; Chiles et al., 1987) and transferred at 299 mAmps for 18 h to a
nitrocellulose membrane (Chiles et al., 1987). The nitrocellulose membrane was blocked
with 5% non-fat dry milk (NFDM) at room temperature for 1.5 h, rinsed with
TBS/TWEEN (30 mM Tris base, 150 mM NaCl, 0.1% Tween 20, pH 7.6), then incubated
with primary antibody prepared in TBS/TWEEN for 1-2 h at room temperature. Two
quick rinses, one 15 min rinse, and two 5 min rinses with TBS/TWEEN were followed by
incubation of the nitrocellulose in the secondary antibody, prepared in TBS/TWEEN, for
1-2 h at room temperature. After washing nitrocellulose six times for 5 min each, the blot
was incubated in 6 ml of a 1:1 mixture of the Enhanced Chemiluminescence (ECL)
reagents (Pierce, Rockford, IL) for 1 min, drained, wrapped in plastic, and immediately
exposed to film.


59


109


115